Ask Slashdot: If You Could Assemble a "FrankenOS" What Parts Would You Use?

rnws writes: While commenting about log-structured file systems in relation to flash SSDs, I referenced Digital's Spiralog [pdf], released for OpenVMS in 1996. This got me thinking about how VMS to this day has some of, if not the best storage clustering (still) in use today. Many operating systems have come and gone over the years, particularly from the minicomputer era, and each usually had something unique it did really well. If you could stitch together your ideal OS, then which "body parts" would you use from today and reanimate from the past? I'd probably start with VMS's storage system, MPE's print handling, OS/2's Workplace Shell, AS/400's hardware abstraction and GNU's Bash shell. What would you choose?

Comments

What?

[Anonymous Coward]

What are these parts I keep hearing about? I use systemd.

Re:What?

[smittyoneeach]

"Systemd is a relatively useful inferior process," said Emacs.

Re: What?

[Anonymous Coward]

Just needed the comma, for dramatic pause plus *puts on sunglasses on* ;)

Re:What?

[plopez]

systemd is too immature. I am waiting for systemv

Re:What?

[Anonymous Coward]

We Scandinavians are eagerly waiting for the SystemÖ. There is even a saying that something should be put to the ö-folder.. On the subject, heterogenous distributed computer over all the machines and devices within the network delay, Amoeba style but evolved.

Re:What?

[Anonymous Coward]

Nah, just take the worst code from Windows ME, Vista, 7 and 8.0, blend until the lumps are not visible from space, and dub it "new" a.k.a. Windows 10.
Systemd at its worst cannot hope to ever compete against Microsoft.

Re:What?

[Anonymous Coward]

Anyone going to bite on that weak sauce?

Re:What?

[Anonymous Coward]

systemd is too immature. I am waiting for systemv

System V was introduced in 1983.
see: https://en.wikipedia.org/wiki/UNIX_System_V

Re:What?

[Anonymous Coward]

*whoosh*

Re:What?

[thinkwaitfast]

BeOS was realtime. QNX/Photon could also do this. The "problem" with this is that you are trading responsiveness for throughput. In order to never see a glitch in the GUI, your server performance drops some measurable percentage and that doesn't look good on benchmarks.

Re:What?

[m.alessandrini]

D > V (at least in roman numbers)

Re: What?

[Anonymous Coward]

That will not stop it from trying!

Re:What?

[Anonymous Coward]

THBBFT! Baremetal OS has everything you need.

Re:What?

[John Allsup]

Then we will await systemmic!

Re:What?

[Hognoxious]

How's the Greek decision going down over there in Berlin?

Re:What?

[blind biker]

Actually, BeOS was near-realtime. QNX is true realtime.

Re:What?

[Anonymous Coward]

*meta whoosh*

Re: What?

[ihtoit]

Just needed the comma, for dramatic *puts on sunglasses before heading off to the kitchen to dress a turkey and returning with a batch of freshly squeezed raisin juice* ;)

PAUSE!

FTFY.

Re:What?

[ihtoit]

I'll wait for SystemOfADown.

I'll be here all week.

Re:What?

[ihtoit]

funny you should ask that, a Greek friend of mine just popped up and said that his local store is now accepting Bulgarian Lev.

That's left me in fucking stitches!

(before you take the piss, the Lev is tied to the Euro).

Re:What?

[ArmoredDragon]

I must be a new minority or something. At first I followed all of the arguments against SystemD, but just for shits and grins I tried Ubuntu Server 15.04 in its default configuration when I built three new network appliance VMs and...I actually like the result. I never did figure out how to get upstart to reliably make e.g. rtorrent restart when it crashes (and it crashes a lot) with upstart, whereas with systemd its crash recovery seems flawless, and it was easy to configure (you just need one line.) It was also very easy to tell it to wait until NFS was actually accessible before starting (upstart only gives you the ability to check that the NFS service is running, not that the NFS partitions actually mounted first.) With upstart I had to hack together a bash script that checked if NFS was mounted and then launch rtorrent as just a regular (non service) process.

Not only that but I was also impressed by how fast the reboots happened as I was setting up these VMs. Namely, I could open a new putty session to the server immediately after I issued the reboot command and closed the last one, whereas with upstart I would wait at least a minute or so. Now I no longer feel tempted to keep the vsphere client open just to watch its progress either.

Of course, I'm not a Linux guru as this server is mainly working as hobby tool and as a lab tool for running Cisco CSR1000v's for my CCIE training (though the server ONLY runs Linux, vSphere 6.0, and Cisco IOS, so I guess unlike most power users I have no dependence on Windows.) So having said that, I'll still defer to somebody who is a Linux guru for better judgement on systemd.

Re:What?

[DanJ_UK]

That's fucking priceless.

Re:What?

[alva_edison]

Then we will await systemmic!

Is there some significance to 1099, or is the joke related to the word mic?

Re:What?

[TarpaKungs]

"Systemd is a relatively useful inferior process," said Emacs.

I'm surprised noone has used init=/bin/emacs

Re:What?

[Agripa]

"The essence of superiority is delegation; I have therefore delegated these tasks to Systemd." - Emacs.

Re:What?

[Anonymous Coward]

*crickets* You're not going to get many laughs around here by mentioning passe shit bands.

Also, your sig is supposed to be funny, but you don't seem to know how to pronounce Chopin. It's "show-pan", not "choppin" or "shoppin". So "Gone show-pan, have list, bock in a min-yoo-et" is what your sig says.

Re:What?

[ihtoit]

the fuck did I just read?

Re:What?

[Bugamn]

Since you asked: http://www.informatimago.com/l... Is this what you where looking for?

Re: What?

[Anonymous Coward]

Unlike the "new" drachma which will be worthless!

easy

[Anonymous Coward]

Revive BEos and let it work with win, nix, mac and android files out of the box.

Re:easy

[Zobeid]

Just, for goodness sake, don't require all app development in C++!

Re:easy

[epyT-R]

No thanks. As it is, today's systems suffer from too much dependency on non-native bloatware.

Re:Easy

[lucm]

Solaris networking is the worst of the worst. Servers ship with the same MAC address on all network adapters, and they get confused by port speed autonegotiation. Those are problems that even Windows Me didn't have.

Re:Easy

[Anonymous Coward]

FreeBSD kernel, Solaris networking/clustering capabilities, and a sort-of Windows UI (Imagine Windows 7, not the metro interface)

dTrace, up-to-date ZFS, and QFS (for a really FAST file system) from Solaris, too.

(ZFS has a lot of great features. Speed definitely ain't one of them.)

Re:Easy

[rev0lt]

Well, you seem to be way more versed in it than I am. I've used both Solaris Express and OpenSolaris on x86-64, without ever having this problem (I did some simple cluster configurations with both Solaris-HA and Sun Grid Engine and I'd guess this would be a major issue). Their virtual networking features were ahead of their time (at least UNIX-wise, I do know VMS, but mostly from reputation - typing in commands on a 1985 machine doesn't really count as experience), but maybe the guys buying actual SPARC boxes had their own issues.

Re: easy

[Anonymous Coward]

I was thinking of beos as a good starting place. It had a lot of what I liked already.

Re:Easy

[ArchieBunker]

What hardware had the same MAC on every adapter?

Re:easy

[Anonymous Coward]

No thanks. As it is, today's systems suffer from too much dependency on non-native bloatware.

Oh for fuck sake stop pretending like everything today sucks and the "back in my day" systems were so much better. Those systems were slow, buggy, unstable junk that was cobbled together using various different vendor-specific implementations. Remember 3D graphics of the 90s? It was junk, we needed unified abstractions to avoid the instability of those old architectures.

Yes if everybody promised to play nice you wouldnt need to worry about the overhead of sandboxing applications or about malware and various other security tactics but this overhead exists as personal computing has become more ubiquitous. And if nobody cared about information presentation you could have the static web of the 90s too. But there's a subset of old people like yourself that can't adapt to new technology so just criticize it in order to try to stay relevant.

Re: easy

[Anonymous Coward]

Exactly who this question is aimed at. Nothing is perfect. Where is the good part of the meatloaf we all have to eat at the Internet cafe? Tell me what you belive my CIO should hear when someone makes a recommendation.

Re:Easy

[Anonymous Coward]

IIRC, the pizza box (SparcStation) adapters kept MAC in NVRAM, and if it got wiped or battery died, it defaulted to MAC.

--sf

Re:Easy

[lucm]

Netra (SparcStation, Ultra, etc.) among others had this "feature". You were expected to change it if you needed those adapters to work on the same network, and of course this had to be done with the shitty LOM connection, which itself was a huge pain in the ass.

I'd rather take a job at Best Buy running Windows Update and antivirus scans on cheap HP laptops all day than touch another Solaris machine.

 

Re:Easy

[mdhoover]

That was a "convenience" on old sparc boxes (think boot net installs) for when you didn't know the mac addresses of your physical cards.
Easily turned off in openboot by setting local-mac-address? = true.

As for networking nowadays, compare the feature set crossbow provides to anything any other vendor provides.

Re:Easy

[lucm]

What exactly is offered by Solaris crossbow that doesn't exist in a similar fashion in the IBM (LPAR) or HP (VPAR) virtualization technologies? Nothing but fluff.
 

Re: easy

[Anonymous Coward]

Mmmmmm... Meatloaf.
Sorry, what are you going on about now? All I see is the text equivalent of insanity.

Re:easy

[Anonymous Coward]

My 3D graphics in the 90s were awesome. Dual Voodoo 2 cards and paired with a Matrox Millennium II. Maybe you just had a shit computer.

Re:Easy

[l0n3s0m3phr34k]

just a few days ago there where MANY stories on here about Solaris networking costing businesses thousands of dollars from all these...bugs...

Re:Easy

[Anonymous Coward]

That hasn't been true since the 1990s. Also, strangely enough, it was standards conformant. Somewhere it was decreed that a single host would have one MAC, and they actually did it.

Re:easy

[Pseudonym]

Yup, for the UI you want something more like QNX Photon.

Re: easy

[Anonymous Coward]

I second that. And massive RAM amount.

Re:Easy

[Anonymous Coward]

Haha, you almost got me there "Those are problems that even Windows Me didn't have."

Re:Easy

[Anonymous Coward]

OSX that is, then.

Re:easy

[Anonymous Coward]

Non-native bloatware is somehow better than native bloatware? And how does one become dependent on non-native vs native bloatware? And what 'bloatware' are you talking about that all systems are somehow dependent upon? (I assume you mean dependency 'to function' and not some dependency in the chemical sense.) Or perhaps you're the type that misses 1 floppy versions of wordstar and basica and anything larger than 360KB is considered 'bloatware'?

Then again I've read your other posts and it appears you're a mysoginistic idiot with dreams of high framerates and right-click reflexes, so i'll count this the statement of an idiot and move on.

Re: easy

[Anonymous Coward]

It's called Haiku. Look it up. https://en.m.wikipedia.org/wiki/Haiku_(operating_system)

Re: easy

[Anonymous Coward]

2 out of 3 aint bad.

Re: easy

[Anonymous Coward]

found the fatty.

Re:easy

[shaitand]

There is quite a bit that is legitimate in criticizing the new trend of 400 layers of abstraction each adding their overhead for the sake of rapid development. Those old development models produced more stable and dramatically more efficient software. Nobody really denies that, they just argue that developer time is more valuable than computer time and improvements in hardware make up the difference in most use cases.

But what happens when hardware stops getting dramatically faster? We'll have to go back to making things more efficient to see gains and suddenly you'll have a great deal of respect for dad who could make a word processor with 90% of the functionality of word that weighed in under 1mb and ran smoothly on as little ram with a processor your smart phone could emulate 200 times over. There is a reason where the browser makers, the home of high level abstraction and high level languages, are finally all actually in agreement and collaborating and the thing that brings them together is making c/c++ compile to a uniform standard that all browsers can execute at near native performance.

The circle comes around, it always comes around.

Re:easy

[Anonymous Coward]

No fucking way. BeOS had a much better UI. In fact, it still has a better UI than any operating system available today.

Wouldn't mind have a QNX kernel running the BeOS UI though.

Re:easy

[Anonymous Coward]

"they just argue that ... improvements in hardware make up the difference in most use cases."

Except hardware hasn't really gotten significantly faster(*) in the last decade, just more parallelised (and there _is_ a difference)

(*) Other than storage speed, but that wasn't particularly difficult given the status quo was electromechanical.

Yes, memory is cheaper, so you can throw more in and keep more off the disks but the days of saying "we'll make it up in faster hardware" have been gone for a while. Developer time optimising procedures pays big dividends. (it did anyway, but now it's the only way to make significant performance gains)

Please insert Multics subthread here.

[jeffb (2.718)]

I'll freely admit that I was too much of a newbie to really appreciate Multics during the precious year or two I had access to it, but the single abstraction for memory and files seemed like a great approach...

Re:Please insert Multics subthread here.

[chipschap]

Honeywell got hold of Multics and, being Honeywell, killed it in favor of selling their own super-crappy GCOS (which had exactly one good features, scatter-gather I/O). Then Prime Computer made PRIMOS which they called a mini-multics (very mini).

Multics was amazing for its time, maybe for all time.

Re:Please insert Multics subthread here.

[DrLlama]

There was a desktop OS called Domain/OS from Apollo Systems. Rumour had it that Apollo was founded by Multicians who fled from Honneywell. It was a great OS on a lot of levels, not least you had native Domain/OS, BSD4.3 and System5 UNIX, an amazing shared filesystem, and networking that was literally plug and play.

Then of course HP bought it and killed it in favor of HP/UX, sigh.

Re:Please insert Multics subthread here.

[Anonymous Coward]

GCOS had many faults but was certainly competitive with the other mainframe OSes, such as MVS.

Honeywell also had their own "mini-multics" called MOD-600 on their Level 6 minis. It wasn't very successful either.

But Multics was indeed way ahead. Some of the features it introduced - e.g. dynamic linking - have arguably never been implemented as well since.

Re:Please insert Multics subthread here.

[_merlin]

AS/400 maps all storage (including memory and disks) into a single 128-bit address space. You still end up with a single abstraction for memory and files, approached from a completely different direction.

Re:Please insert Multics subthread here.

[Pseudonym]

Unfortunately for Multics oldbies, the x86-64 architecture has no hardware segmentation in long mode. So you're not going to see the likes of Multics again.

Of course, address spaces are so big these days that using virtual memory to do file I/O is practical.

Re:Please insert Multics subthread here.

[John Allsup]

It has occurred to me before that a 64bit memory space is adequate for most purposes (except really big computers), and then a hierarchical arrangement of 64bit address spaces (essentially give each 64bit memory space its own index, with 0 meaning local-default or something like that). Similarly, I started thinking through the idea of a really minimal Forth for 64bit chips where, rather than strings naming words, we just limit things to strings of 8 1-byte characters, possibly using a different character set than ASCII, and ordering it so that lexicographic ordering is identical to the native integer ordering of the processor (then you can remove all need for string processing from the Forth, and move it, if needed, into a frontend).

Chuck Moore (of Forth fame) once commented about the database. Casual users forget that their filesystem is a database, and quite possibly one that is far from optimal for what they are doing with it. Viewing memory and storage as a database of 4K pages, and working up is something that to a certain extent already happens, but making it explicit in system architecture makes a lot of sense to me.

Ideally I would have a system where a minimal Forth like this sits at the bottom level, then there is a minimal Lisp for higher level stuff, with access to a sufficient compiler infrastructure to do basic assembly and writing of components. As for system architecture, I would strip away many of the abstraction layers we currently have. View both storage and memory as a database of 4K pages, each of which is natively viewed as a vector of 512 64bit integers, and viewing smaller cells as fractions of a 64bit integer, and larger cells as vectors of 64bit integers.

As one goes up, implement systems for the various programming paradigms, borrowing from languages such as the various oo languages, the various functional languages, and things like erlang, and the bright ideas that turn up in webserver systems such as nginx and node.js (that is, lightweight 'threads').

The difficulty with any dream like the above, however, is to avoid the 'design by committee' problem of ramming in every pretty idea you see, and of trying to arrange things so that the complexity of the end result system does not blow up massively as things grow towards what we would expect of a modern operating system.

Re:Please insert Multics subthread here.

[Anonymous Coward]

Holy teletype font.

Whatever the fuck you're doing, stop it.

Re:Please insert Multics subthread here.

[Old Bitsmasher]

Just to set the record straight, after the GE merger it was the people who were originally from Honeywell who fought to save Multics and the GE alums in Phoenix (AKA the Dinosaur Factory) who fought to maintain the dominance of God's Chosen Operating System. I know this because I was there, on the losing side.

Re:Please insert Multics subthread here.

[Dynamoo]

Agreed, GCOS is horrific and it is the OS That Will Not Die.

Multics (RIP) had some cool and interesting features. Everything was a file. Security Rings. A very consistent set of commands. But at a couple of million bucks at least per installation in 1980s money, it was a fairly niche product. I do remember having arguments with Unix folks about the strengths and weaknesses of both OSes. Of course Unix and its derivatives are absolutely everywhere now.

Re: Please insert Multics subthread here.

[Anonymous Coward]

The whole idea of mmus protecting from malware and bugs is badly wrong.

Burroughs icl and elbrus were much better until the bell labs stuff was foisted "for free" on the globe.

Memory safe languages for kernels are the way foward to smoke out the cyber war domain.

Mozilla and google already lead the way out of this he'll with rust and swift. Burn c and unix on the stakes !

Re:Please insert Multics subthread here.

[400_guru]

Yes! IBM i (The current generation of the AS/400) with its 128 bit address space would work well in a computer with only memory, no disk! In fact it was envisioned that way at its inception. Additionally the abstraction that it brings allows program objects going all the way back to the System/38 to run on current POWER8 hardware unchanged. That's software investment protection! Of course why you would want to run 30 year old software is a valid question but you could if it was valid and needed. The second benefit of the abstraction layer is that new hardware is easily introduced and over it's history OS/400 (the O/S that ran on the AS/400) and now IBM i have hit many first in the industry points. These include memory and storage technologies such as the first SAS based SSDs. Oft overlooked except by sysadmins the command line of IBM i is also second to none being very intuitive, very consistent, and easy to learn. One example is vfytcpcmn which does the same thing as ping. (Yes ping works too!).

Re:Please insert Multics subthread here.

[Anonymous Coward]

Most purposes? Really big computers?

Uhmmm ... in what reality is a 64 bit address space not big enough? That's
18,446,744,073,709,551,616 bytes!
or 18,446,744 GB of physical RAM per CPU (18,446 TB ... or enough to address every byte in over 184 RoadRunner super-computers).

For scalability, you don't share all the RAM of multi-CPU machine. Read up on NUMA architecture. Your address space need only be per node.

In what universe is 64 bit not big enough?

And FORTH and LISP built into the OS? Sure ... let's add a Smalltalk layer and Atari BASIC while we're throwing in useless shit nobody will ever use. How is the language part of the OS?

Re:Please insert Multics subthread here.

[Anonymous Coward]

Back in the dark ages of the early 1980's I sold QNX-based PC's to USGS field researchers. They wanted Multics, but couldn't fit one into their suitcase! The more important thing was real-time and a standard C compiler, which they got with QNX (QUNIX back then until Bell Labs, owner of Unix, complained about the name).

Throw it all out

[Anonymous Coward]

"We have a whole valley full of people talking UNIX versus MS-DOS. What do you need any of that for? Just throw it all out; get rid of all that nonsense. Maybe you need it for computer scientists, but for people who want to get something done, no. Do you need an operating system? No."
  - Jef Raskin

Re:Throw it all out

[epyT-R]

So how would you do away with the operating system and still enable people to 'get something done'?

Re:Throw it all out

[randalware]

I think Raskin was talking about one of his projects.

Swift card Canon Cat

You just start typing and the computer figures out what your wanting to do.
opens the word processor, spreadsheet, etc

Re:Throw it all out

[Anonymous Coward]

insert removable storage, turn computer on

Re:Throw it all out

[epyT-R]

Right. The software handling such input would be the OS, or at least its shell. We had that already too, it was called 'Active Desktop' and it sucked horribly. Today, they call it 'semantic desktop' and it still sucks. For whatever reason, MS, apple, google, and others keep trying to bring it back, and make it stick where it doesn't belong.

Re:Throw it all out

[epyT-R]

something to be said for that, but I think most people would see running one program at a time as getting in the way of productivity. Personally, I do miss the simplicity of those days.

Re:Throw it all out

[Anonymous Coward]

This is actually what I would like... a virtual machine on each of my SD cards that runs on whatever device I plug it into, and which disappears from the device when I remove the card.

Re:Throw it all out

[chipschap]

You mean like on Android. (Effectively. Yes, it's not the same, but in user space with few exceptions that's a big limitation of tablet computing.)

Re:Throw it all out

[Anonymous Coward]

Yeah, he was talking about "information appliances" which would be cheap and let people do a few simple tasks without having to run special software. It's the antithesis of the general-purpose computer, which is so flexible precisely because you can run all kinds of software on it.

Still, I kind of think he was right. An operating system is just another program or set of programs, and lot of the tasks it performs are only necessary because the Von Neumann architecture is so basic. We're starting to see things now like NAS drives with fileserver and network software already installed - you'll never have to mess with drivers and filesystems with those, the hardware/firmware takes on OS functions.

Re:Throw it all out

[OrangeTide]

I sure do miss Jef ;(

He would occasionally show up at SVFIG (Silicon Valley Forth Interest Group), and we've all had fun chats with him.

Re: Throw it all out

[Anonymous Coward]

You know that most NAS stations actually run a OS like a lightweight linux or bsd? ;)

Re:Throw it all out

[mlts]

The closest we came to that was the first NeXT machine which had an optical drive (or perhaps two), and initially, no hard drives.

Application level, we did have U3 for a bit, which would do similar... run a program, then it would exit when the USB flash drive was yanked.

I wouldn't mind some form of hypervisor where the desktop OS would be like ESXi, except with the ability to use the console for graphics with VMs. This way, it is a lot harder for a rogue program to jump from a client OS to hardware.

Re: Throw it all out

[TheRaven64]

They do now, but mostly because that's what's cheap. For high-end stuff, OS bypass is the buzzword of the day and as more of that stuff starts to be moved into userspace it becomes plausible to have a tiny OS and get rid of a lot of the complexity of something like a UNIX kernel ('lighweight linux or BSD' is amusing, given that the kernel of either is several MBs).

Re:Throw it all out

[John Allsup]

One thing I love about xfce (I tend to use UbuntuStudio on most of my machines) is the ability to assign arbitrary keystrokes to commands in the desktop preferences. I thus use the Windows key (call it Meta if you like, but I am too used to calling it Windows, especially since it has a Windows logo on) so that Win-W launches my default word processor, and then if necessary modifiers like shift and ctrl will launch an alternative word processor. Win-T launches the terminal, Win-G does gimp, Win-Shift-G does krita, Win-E does the file manager (having got used to this one on Windows), Win-K does kodi, and so on.
Rather than rely on some dodgy AI to figure out what I want, give the user an easy way to teach it what you want, so that one can easily teach the user how to train the computer to respond. The idea that a few software engineers in a country a few thousand miles away can anticipate how I would want to use the computer is just silly, but that is what Apple, M$, and many desktop environments effectively do.

This hotkey arrangement means that I can go from off to word processor in three touches (power button, which boots to desktop*, Win-W) and about ten seconds. With windows 8, already booted and logged in, it usually takes me longer than that, and many more touches, to get a working word processor. Likewise for other common apps (Win-B does Firefox, Win-Shift-B does Chromium)..

One thing I dislike about how things are on my Linux boxes at present is a kind of 'fragmentation' in the sense that only certain bits of my environment can be directly controlled from the command line, and the command line is limited compared to a language such as Python. I would personally take the idea of sending messages in a command language that would, in practice, resemble a simplistic Lisp, as a core of all applications. UI events like changing font in a word processor or changing colour in a paint program would resolve to commands sent to a mailbox in the app somewhere. This would make the MVC paradigm much more explicit, and I would have the ability to 'log in' to a command processor built into every program (essentially stick such things into the runtime -- look at F-Script anywhere on Mac Os X for an idea about what I am describing).

In particular, if I can conceive of a simple task for, say, my word processor, I would like there to be an easy method to tell it to do this. I am a big fan of the idea of pervasive scriptability, But that does require more thought about the nature of the scripting languages: you want something data-structure based (like Lisp) that is easy to compose programmatically (so that any old scripting language can compose a script for this pervasive scripting language) and easy to pass between processes, whether local or remote (so that a script can throw a complex request to a process somewhere else to handle).

*and I take care to use encrypted containers for anything that I don't want someone with physical access to my machine to get at.

Re:Throw it all out

[ihtoit]

I already use such a system on my home terminals. Everybody has their own SD card (a key, if you like) that contains their OS-of-choice and workspace. No matter which system they plug into, their dekstop and settings look identical and even if they were plugged in remotely, if there's an internet connection available they'll connect to the home shares. The exception is my laptop and my netbook, both of which have internal persistence (hard drives) necessarily because I also administrate the file services through them.

Re: Throw it all out

[Anonymous Coward]

Yes, but my point is that the offloading of responsibilities from the PC's OS could eventually lead to most of the "OS" being unnecessary!

Re:Throw it all out

[Grishnakh]

So how would you do away with the operating system and still enable people to 'get something done'?

That's easy: you go back to the days of MS-DOS (which was nothing more than a program loader). We got plenty of stuff done back then, it was just slower and crappier because changing tasks took so long: you had to save and exit your application, then start up the other application and load your work. If you had to constantly switch between two things, it was a real PITA.

But you could "get something done". Countless secretaries got lots of work done in WordPerfect in those days.

Re:Throw it all out

[cbhacking]

Little-known fact about Windows: you can have it do keyboard shortcuts like that too! This isn't even new; I know it was in Windows 2000 and is probably even older. The only problem is that it can't replace built-in or app-defined shortcuts, so things like Win+W won't work (On Win8, at least, that's a Search panel for Settings).

Right-click any shortcut (including from the Taskbar or Start menu/screen), and select Properties (or open Properties some other way). There will be an option for "Shortcut Key". Select the option, press the combination of meta-keys + character to use to launch the shortcut, and hit OK.

Most of...

[Anonymous Coward]

Most of QNX.
Which is why.....

MarkT

Re:Most of...

[OrangeTide]

The main problem I have with QNX is with QSSL (now Blackberry). The royalties for QNX have always been high, and has made it tough for me to convince management to use it in projects. The technology is fantastic though.

VMS queue manager and VMS breakin evasion

[Anonymous Coward]

I would love to see the VMS queue manager on Linux so we could have proper batch job support and print queues with real forms management.

I would also like to see VMS style breakin evasion as part of the core Linux operating system instead of various third party addons

Re:VMS queue manager and VMS breakin evasion

[Anonymous Coward]

To followup on this, there also things the older operating systems could learn from today's operating systems.

For example, with VMS, I would like modular plugin filesystems, and it would have been nice if the kernel had been designed with porting in mind.

DCL (the VMS CLI) could also do with a major rework and enhancements.

I'm also in favour of microkernels as a general good idea so that we don't have massive chunks of privileged kernel mode code all in the same address space and with no internal isolation between logically distinct code.

Re:VMS queue manager and VMS breakin evasion

[Nkwe]

DCL (the VMS CLI) could also do with a major rework and enhancements.

True, but there are elements of DCL that are worth bringing forward as well - particularly the ability to define the command syntax at operating system (shell) level and bind it to programs. For those who who haven't developed under VMS, there is operating system (shell) defined syntax with which you declare what parameters, switches, options, etc. that your program desires. The syntax is robust enough to specify which options are optional, required, mutually exclusive, etc. When you build your program, you "compile" in your command syntax and at run time DCL handles syntax checking for you. Coding work for processing command line parameters is greatly reduced. You also get bonus stuff like integration to the help system and automatic shortening of non ambiguous switches.

Re:VMS queue manager and VMS breakin evasion

[Anonymous Coward]

Why would you taint VMS with filesystems that aren't inherently built for clustering and contain all of that bloat? Specifically, what do you find wrong with VMS (other than dealing with USB thumb drives)?

Re:VMS queue manager and VMS breakin evasion

[Anonymous Coward]

AS/400 for simulor for its commands and program calls. Nice is the editor can also access it so building a new "script" you can prompt and fill in the blanks.

Re:VMS queue manager and VMS breakin evasion

[Anonymous Coward]

Interesting - Windows PowerShell does exactly this if you're building powershell cmdlets.

Re:VMS queue manager and VMS breakin evasion

[Anonymous Coward]

The filesystem structures themselves, including ODS-2, aren't built for clustering.

It's the DLM (and the use of the DLM in the filesystem code) which gives the filesystem a clustering ability.

As for what's wrong with VMS, I want the ability to process modern filesystems and hardware (as you apparently do as well with your desire to support USB thumb drives).

Re:VMS queue manager and VMS breakin evasion

[TarpaKungs]

And if you did not want to load the parse tree into your DCL instance at login time (because it would CoW the affected memory pages with an entire new parse tree for the whole of DCL IIRC - it was incredibly slow) you could link it to your program, call your program using the "unix like" alias method then have your main function call the same command line parser libs that DCL used. Bit more like the getopt() model but 100% consistent with the regular DCL experience.

Re:VMS queue manager and VMS breakin evasion

[TarpaKungs]

You also get bonus stuff like integration to the help system and automatic shortening of non ambiguous switches.

I did like that - nice consistency - unlike the bastardisation that is every linux command.

I really hate commands that respond to "-h" with "Type --help for help". FFS - you parsed "-h" - how hard would it have been for you to just link that to the "--help" code!!! Or use getopt and it's more or less for free...

Duh

[sootman]

Windows kernel, Linux UI.

Re:Duh

[drinkypoo]

Windows kernel, Linux UI.

We had that, it was called 'Windows for Workgroups'

Well, it was a UNIX UI. But close enough.

Re:Duh

[whoever57]

Windows kernel, Linux UI.

You actually want BSODs? Or was that a whooshing sound that I just heard?

Re:Duh

[JWSmythe]

I'm pretty sure that was a high altitude joke sailing by at mach 1.

Re:Duh

[Anonymous Coward]

The BSOD were pretty much eliminated in Windows NT 4, if not earlier. Yes, NT 4. The BSODs that people typically got were inevitably the result of a badly-written piece of software running in kernel mode. Applications were free to crash as often as they wanted and anytime they wanted and the OS would continue humming.

That Unixes get such praise for being stable is largely due to the fact that, early on, they only supported the vendors' unique hardware profiles.

Re:Duh

[Anonymous Coward]

MS changed the BSOD to a reboot. You can get the BSOD back instead of reboot somewhere in the settings. I ditched MS for Linux 10 years ago and will never go back to that bloated/slow POS they call an operating system.

Re:Duh

[lucm]

If you've been away from Windows since before Vista, your opinion is irrelevant. You missed all the fun.

Re:Duh

[ArcadeMan]

Microsoft Windows 95 kernel, Mac OS classic UI, Nintendo Gameboy CPU and display, Radio Shack TRS-80 MC-10 keyboard.

Re:Duh

[Anonymous Coward]

already possible. see kde for windows.

Re:Duh

[MobileTatsu-NJG]

You actually want BSODs?

Who said anything about Win9x?

Re:Duh

[Anonymous Coward]

Windows ME kernel, Squeak UI, U880 CPU, and the keyboard from the Robotron Z1013.

Re:Duh

[MobileTatsu-NJG]

You can get the BSOD back instead of reboot somewhere in the settings.

Even in 2005 you'd be waiting a long time to actually see either happen.

I ditched MS for Linux 10 years ago and will never go back to that bloated/slow POS they call an operating system.

Amusingly 'unstable' is not on that list.

Re:Duh

[chipschap]

Getting a little off topic here, but that was exactly my case .... I was on XP and for the little non-Linux computing I had to do it was fine. I hadn't bought any hardware in years and just last week got an Asus UX-305. It came with Windows 8.1 and I think it would have driven me to jump off the balcony of our 30th floor condo had I not, after 20 minutes, wiped the SSD and installed Linux Mint 17.2.

So maybe my opinion is irrelevant, and maybe I missed all the fun, but I'm happy to continue permanently to miss out on that sort of fun ... and ... you know, actually get something done without being driven to self-destruction.

Re:Duh

[Topwiz]

I actually got a BSOD yesterday on my Windows 8.1 laptop. I also have a Windows 7 computer and have gotten a BSOD doing the same thing. I connected to my work computer using TeamViewer and then on my work computer (Windows 7) I was running vmware vsphere client to check on one of our virtual Windows servers. I double clicked somewhere and all of a sudden BSOD. After the laptop finishes rebooting, I connect to my work computer again through TeamViewer and everything is exactly like I left it. I'm not sure what it is I did to trigger it.

Re:Duh

[Anonymous Coward]

It was not awesome

Re:Duh

[lucm]

You find life unbearable on a Zenbook running Windows 8.1, and you say Ubuntu Jr makes things easier? Unlikely.

What obviously happened is that you got your panties in a bunch when you saw that metro screen and weird start button, and you couldn't be bothered to spend 5 minutes to read or watch a Windows 8.1 tutorial to learn about all the nice features. Instead you blamed Microsoft and went back to your zone of comfort.

It's ok to prefer Linux; I use Fedora on my desktop. But you didn't even give a chance to Windows, so don't come bitching about that O/S and be a drama queen on top of it. Millions of people use Windows on a daily basis on hardware that is a lot worse than your Zenbook and they don't jump from the balcony 20 minutes after powering up their machine for the first time. It's people like you that make Linux users look like petulant idiots.

Re:Duh

[brantondaveperson]

You're joking, but:

Windows threading and synchronisation primitives
Windows memory management, and flexible swapfile (linux's swap partition is bonkers).
Linux low-overhead processes and fast filesystems
OS X use of launchd, and its UI

Re:Duh

[Dr_Barnowl]

You can do Linux swap on a regular file, but using a file is the real bonkers. Files exist in the file system, which means that they don't necessarily represent a contiguous area of disk space, which is the main property you want from a swap file (on spinning rust, anyway).

On Windows if your swapfile grows it may fragment. On Linux the swap is a partition and doesn't grow on demand. If you really need to, you can create a file in /tmp and attach it as swap, dynamically, and remove it when you're done (or just reboot, and pow, it's gone).

All in all, the Linux way makes more sense to me. The Windows swapfile was always a massive pain in the ass, stuck on your C partition eating valuable space and growing unexpectedly when you had an unanticipated memory demand and then never shrinking again.

Re:Duh

[Anonymous Coward]

Hey, KDE and GTK libraries are working on Windows so its only a matter of time before we get that "harmonious whole and finely crafted experience" on Windows. There once was a myriad of shells available for Windows, some of them really interesting.

Re:Duh

[brantondaveperson]

I don't think that there's any evidence that the linux swapfile performs better - and in any case why would it being unfragmented be an advantage? Memory access is random, and so swapfile access is random, and so why does having it non-contiguous cause an issue? Added to which, SSDs are becoming much more widespread, meaning the fragmentation issue vanishes.

In any case, can you make Linux use a swapfile permanently? One system I look after needs more swap, and I really don't want to repartition the entire drive just to increase the available virtual memory.

Re:Duh

[tlhIngan]

and flexible swapfile (linux's swap partition is bonkers).

Uh, you could use a swap file with Linux for many versions now. I know I had it running on a Linux machine way back in 1997. I can't remember why we did it (I think the hard drive with the swap partition died and we needed the machine back up for other services), but it was documented including all the steps from using dd to create the swapfile (it needed a pre-allocated file), to mkswap and then swapon.

In fact, the mkswap man page tells you how to make a swapfile. And I know that man page was virtually identical back in 1997.

Re:Duh

[brantondaveperson]

Well yeah, but it's still a pain, and requires a contiguous file, and multiple commands to set up, and isn't the default.

So I'm sticking with - Windows-like swapfile support.

Re:Duh

[Anonymous Coward]

OS X use of launchd, and its UI

We already have that, it's called systemd

Re:Duh

[dcollins117]

In any case, can you make Linux use a swapfile permanently?

Yes, you just add a line to fstab to make it permanent. There's a brazillian guides on Google that will show you how.

Re:Duh

[c0l0]

One system I look after needs more swap, and I really don't want to repartition the entire drive just to increase the available virtual memory.

Sure. You can reference swap files the very same way you reference swap partitions in /etc/fstab, and also have more than one swap-file/partition in use at a time. Just create an empty, non-sparse file using, for example, `dd if=/dev/zero of=/SWAP bs=1M count=1024` (this exact invocation will make it 1G in size), run `mkswap` against it, and add that resulting file's pathname to your fstab the same way you did with your swap partitions' device node. Run `swapon -a` to activate it for this uptime - done.

Re:Duh

[Demonoid-Penguin]

Sure. You can reference swap files the very same way you reference swap partitions in /etc/fstab, and also have more than one swap-file/partition in use at a time. Just create an empty, non-sparse file using, for example, `dd if=/dev/zero of=/SWAP bs=1M count=1024` (this exact invocation will make it 1G in size), run `mkswap` against it, and add that resulting file's pathname to your fstab the same way you did with your swap partitions' device node. Run `swapon -a` to activate it for this uptime - done.

That enables swap - it doesn't mean it'll ever be used. Which is what I guess the OP is asking (a Linux sysadmin without a clue).

First, as you've suggested create a swap partition or a swap file (for performance put one on every non-RAIDed drive). Then set swappiness.

Choose a value between 0 and 100 where 0 means the kernel tries to keep everything in RAM and not cached to disk and 100 means it aggressively caches to disk to free RAM. The default in Debian is 60, which is OK but conservative. Laptop users should in any case use a low value to reduce writing to disk (because writing to disk negates benefits of power management and runs down your battery very quickly).

# echo 'vm.swappiness=20' >> /etc/sysctl.conf

If you really want permanent swap (a compelling use case would be best).

# echo 'vm.swappiness=100' >> /etc/sysctl.conf

Fucknose why anyone would actually want that though.

Re:Duh

[jabuzz]

Yeah and using partitions is bonkers as well. LVM is the way to go for anything that does not involve huge (we are talking 100's of TB upwards) clustered file systems.

Re:Duh

[TheRaven64]

Well, except that NT4 had a fun bug with the uptime counter, so if you actually did manage to go around 47 days without a BSoD, you'd get one when the counter overflowed. The fact that it took several years for anyone to discover this bug shows how 'eliminated' BSoDs really were by NT4. Oh, and NT4 moved the graphics drivers back into the kernel (including font rendering, which is why the TTF parsing vulnerabilities found a year or two ago were kernel exploits on Windows), so there was a lot more badly-written software running in kernel mode. I mostly got BSoDs on NT4 from the Soundblaster drivers - Creative Labs should never be allowed near ring 0.

Re:Duh

[TheRaven64]

Windows threading and synchronisation primitives

What windows synchronisation primitive allows:

• Timed wakeup (i.e. try to lock, time out if you fail).
• Adaptive mutex behaviour (spin in userspace for a bit before calling the kernel).
• Can atomically be released when you sleep on a condition variable and reacquired when you wake.

Give up? So did the developers of the Microsoft C++ stack, which is why their std::mutex uses something custom, whereas implementations for POSIX systems just use pthread_mutex.

Re:Duh

[TheRaven64]

OS X users the same underlying functionality from a UNIX-like VM subsystem, but has a dameon that monitors the amount of used swap space and creates new swap files when they're required. This gives you the flexibility of the Windows model, without the complexity in kernel space.

Re:Duh

[Christian Smith]

I don't think that there's any evidence that the linux swapfile performs better - and in any case why would it being unfragmented be an advantage? Memory access is random, and so swapfile access is random, and so why does having it non-contiguous cause an issue? Added to which, SSDs are becoming much more widespread, meaning the fragmentation issue vanishes.

In any case, can you make Linux use a swapfile permanently? One system I look after needs more swap, and I really don't want to repartition the entire drive just to increase the available virtual memory.

dphys-swapfile is available on all distributions that I know of.

Still, swapping from a partition will be faster than swapping from a file. There are simply less layers to traverse, but whether it's a perceptible difference, I just don't know as I've not measured it.

But you're right, SSDs are obviating the need for a contiguous swap file/partition. Swapping is inherently random in nature, which translates to random IO on contiguous or non-contiguous swap.

I have wondered what sort of difference log structured swapping would have? Swapping out would reduce to a contiguous operation as you'd be writing to the end of the log, and swapping in is effectively random IO anyway, so would not be impacted by the fragmentation introduced by the log structured writing. Again, not that much of an issue with SSD, but with HDD, it might be a measurable benefit when trying to free memory.

Re:Duh

[goose-incarnated]

I don't think that there's any evidence that the linux swapfile performs better - and in any case why would it being unfragmented be an advantage? Memory access is random, and so swapfile access is random, and so why does having it non-contiguous cause an issue?

While memory access is indeed random, memory usage is not. Fragmented memory for MxN array accessing would blow chunks because the data is not localised.

Added to which, SSDs are becoming much more widespread, meaning the fragmentation issue vanishes.

I do not know about this - I would hazard a guess ... the issue will still be there, due to caching on the SSD. I imagine pages of fixed size would be cached by SSD controller and non-localisation of your data on SSD would result in multiple cache misses, while localised/non-fragmented data would not.

Re:Duh

[goose-incarnated]

Well yeah, but it's still a pain, and requires a contiguous file, and multiple commands to set up, and isn't the default.

So I'm sticking with - Windows-like swapfile support.

TBH, for 99.999% of computer users it really doesn't matter; the slowdown from a fragmented swap is unlikely to get noticed in most things. For high performance and/or scientific code it will result in a small speed advantage in low memory situations. The fact that partition is ever so slightly better than swapfile for a tiny fraction of users just doesn't matter anymore. Use the swapfile - it's more convenient anyway.

Re:Duh

[Anonymous Coward]

Hi,

The problem is - and my 67 year-old mother, about 10 million businesses, and the nice person you replied to agree - is Windows 8.0/8.1 is an immense paradigm shift, but in a direction which is frankly counter-productive.
It was set up to facilitate touch-screens, Surface Windows PCs, and so on.

Essentially: business (and my 67 year-old mother for that matter) don't want swish new Metro screens, a start menu that makes it actually difficult to locate your MS Office apps, and so on. They want all the improvements, but more importantly: they want to walk in Monday morning after the "upgrade" and to actually be able to work out what the hell they are doing, how to print something, where Notepad has gone, and so on.
They don't WANT to have to spend days and days screwing around trying to find the "right" icons on their Metro screen, and getting rid of the weather, CNN news, travel crap and so on.
They want to just get on with working.

My mother (and I suspect lots of businesses) took her new PC with Windows 8.1 and gave it away to carry on using her old (5 year old) Win XP machine... because all the bells and whistles, plus the fact she just could not find her s*&t meant it took her PC from being a simple "My browser is here... My Euchre program is here" exercise to becoming a "Why in the name of all that's holy do I give a CRAP about Expedia, or CNN, or the weather. I'm next to the window for God's sake!

WHERE'S MY FIREFOX, AND MY BIRTHDAY CARD PRINTING PROGRAM??!!!" exercise.

As a user of Mint/Ubuntustudio, running XFCE I get this completely... especially when I see Ubuntu's current desktop, and think: "WHY do I need to type the name of a program to run it? I WANT MY DESKTOP ICONS BACK!"

Re:Duh

[Anonymous Coward]

Ever seen that huge space-wasting, untabbed file manager in Windows 7? It's what I hate most about it.

Re:Duh

[dbIII]

What obviously happened is that you got your panties in a bunch when you saw that metro screen and weird start button

And then found the controls were hidden off the side of a screen on a touchscreen - yes - unbearable and I had to hand it back before the urge to smash it took over.

Re:Duh

[dbIII]

That enables swap - it doesn't mean it'll ever be used

Then you remove the other swap.

Re:Duh

[Megol]

They did? Well I wouldn't know even though my main computer runs Windows - because it been ages since I've seen a BSOD/reboot.
My current computer never got any - and that one is from 2012. The computer I have before had none, the computer before that one did actually got one - I added a non-compatible MXM card that caused a crash before Windows launched. The computer before that one had no problems, neither did the computer before.

And then I honestly don't know as my memory doesn't stretch back too far.

Re:Duh

[linuxrocks123]

requires a contiguous file

No, it doesn't require a contiguous file -- there's no way in most Linux filesystems to even guarantee a file is contiguous. It will work better with a contiguous file, though, because swap space works better when it's contiguous. That's actually why swap partitions are the default -- all partitions are contiguous. Now, contiguity doesn't matter so much if you're using an SSD for swap, but using an SSD for swap space is retarded anyway because that type of usage pattern on an SSD is likely to kill the drive.

it's still a pain

It's not a pain -- it takes about 20 seconds to do. Maybe an entire minute if it's the first time in your life you've done it and you have to look at the manpage for swapon.

multiple commands to set up

Linux uses the command line for system administration. If the 3-command dd/mkswap/swapon sequence for setting up a swap file is enough to blow your mind, you should probably not try to self-administer a Linux system. Instead, if you want to use Linux, you should get a technically-inclined friend or relative to handle the system administration tasks for you.

I'm not being sarcastic -- I know you probably consider yourself a tech person, but if you find the command line that toxic, being a Linux administrator -- as opposed to user -- isn't something you'd enjoy. The admin/user dichotomy works very well for Linux and is one of two ways non-technical people can use Linux long-term. The other model is "appliance-style"; Android and Chromebook systems are good examples of that model.

I administer Linux systems for multiple friends and relatives, and it takes very little time on my part. Linux people like other people to use Linux. Ask around, and I'm sure you'll find someone willing to help you out.

Re:Duh

[Anonymous Coward]

Modded as a joke but actually isn't. The windows kernel is pretty awesome, its all the shit that is on top that sucks. Its like Gordon Ramsay made a dish and Robert Rauschenberg and Trent Reznor colaborated on the plating and condiments.

Re:Duh

[AmiMoJo]

It's interesting that people seem happy to replace their Linux window manager, or hack around with settings all day to make KDE/Gnome nicer etc. from their default Ubuntu install, with its hooks into commercial search results and the like. Then Microsoft give you the ability to replace the start menu (there is an API for it) so you can customize to your heart's content, and the same people bitch and moan about the default sucking and completely ignore the alternatives.

Re:Duh

[Megol]

I think you are mistaken - Windows 95 did have a timer bug but I can't remember one for NT 4. A quick search doesn't reveal any NT 4 timer bugs.

And I think all would agree that Windows 95 wasn't a stable OS - partially by choice. It was very MSDOS compatible to such degree that not all resources were protected and provoking a BSOD was trivial.

Re:Duh

[John Allsup]

Cue the inevitable cartoon: http://www.toonpool.com/cartoons/Drag%20and%20Drop_22143

Re: Duh

[Entrope]

Being able to replace the default Start menu (or entire initial screen) does not excuse the default being so awful.

Re:Duh

[bingoUV]

It's people like you that make Linux users look like petulant idiots.

Spoilt, actually. UI that changes when the user is ready, even when while receiving the latest performance / security / compatibility updates. Not when the corporate overlords want you to switch UI.

I ran Enlightenment (E 16) in my main system for 9 years - 2003 to 2012. Switched (of my own accord) to fluxbox. Still using it. Using FVWM2 on other systems where it makes sense. Not once did a UI change on my main system without my seeking the change. Same configuration file, moved across upgraded hardware.

On non-main systems, I did experiment with Gnome, KDE which change whimsically. But if I avoid Microsoft's stuff, I can keep things stable where it matters.

Re:Duh

[Demonoid-Penguin]

That enables swap - it doesn't mean it'll ever be used

Then you remove the other swap.

It still doesn't mean it'll ever get used.

Re:Duh

[TheDarkMaster]

Dude, you reversed the polarity of your post :-)

Re:Duh

[Anonymous Coward]

nice features

Dude, even Windows fanboys agree that Win8 is a piece of shit. And it lost customers. Even people who have been Windows users since the late 80s are looking forward to Win10, usually so they can get that crap off their SO's laptop. To paraphrase Agent Smith, how do you watch a Windows 8.1 tutorial when you can't find the video player?

Re:Duh

[Anonymous Coward]

If you think a Windows system is always stable, you've obviously never owned an AMD graphics card.

Re:Duh

[Anonymous Coward]

Check out sprysoftware.com
It is a portable Win32 API for *nixes

Re:Duh

[laffer1]

You never owned an iomega zip drive then. The NT4 kernel would BSOD constantly with an external parallel version of that. I've experienced a BSOD on every OS through Windows 7. In windows 8, it's now a :) so I don't know if that counts. (i've seen it too)

Windows tends to crash now due to hardware problems. Most BSODs in NT are from bad hardware or bad drivers.

Re:Duh

[laffer1]

This can't be accurate. My boss at the time had an NT4 server running SQL Server that he left on for a year. It only had SP1 installed. No security updates.

It was kind of a nightmare to patch when we finally did. Did I mention it was the billing database and that was back when we didn't have the PCI compliance standards of today.

Re: Duh

[Anonymous Coward]

I think the two key things windows forgot with 8.0/8.1 (and Ubuntu for that matter) were discoverability and the advantages of doing stuff by rote.

The old start menu was a wonder of discoverability. Basically if your computer had it, it was in there somewhere, and in a reasonably (if not entirely) ordered fashion - you could explore, discover and learn to your heart's content. The new search option you take a wild guess at a keyword and hope: if you don't recall the name of that cool program you haven't used since last year, too bad, and good luck working out what came pre installed on that nice new laptop.

And for rote learning: do something often enough (eg click start, navigate through menu x to run wowbiz 5.2) and you can do it without thinking. The new start 8.0/8.1 search metro thingy appears designed to prevent that, which is a genuine pain in the neck.

Re:Duh

[CronoCloud]

As a user of Mint/Ubuntustudio, running XFCE I get this completely... especially when I see Ubuntu's current desktop, and think: "WHY do I need to type the name of a program to run it? I WANT MY DESKTOP ICONS BACK!"

Fedora 22, XFCE here. I felt the same way about Gnome 3.

Re:Duh

[Anonymous Coward]

OTOH, it's not his fault the UIX designers of Windows have no taste. Windows 8.1 is godawful, filled with useless eyecandy that will frustrate power users and utterly confuse normal users. I'd say he was pretty tolerant of abuse to endure 20 minutes of pain. Is Windows 8.1 an improvement on Windows 7, or a heavy-handed tactic by marketing gorillas that simply wield too much power over their customers?

IMHO, Apple got it right, but is often too limited. Ubuntu got it even better, and can often be extended if you are feeling brave and looking for more work.
As a forced customer, why waste time and energy? Windows XP, 2000 and 7 were fine, but it does makes sense to stay away when things smell rotten.

Re:Duh

[goarilla]

That's not true, they are not eliminated. There is plenty of crappy kernel mode software out there (drivers, filesystem filters).
And the following (mostly hardware) problems also often cause BSOD's:
- Overheating GPU's.
- Broken stick of RAM, often due to dusty working environment, overheating
or abuse.
- Bad Hard drive causing corrupt system files.
- ACPI Power issues (hibernate,suspend,...).

Now most of these problems do crash all systems.
But personally I find the kernel panic and short stack trace on the console a lot more helpful
than the BSOD (not counting the MEMORY.DMP or Minidump files since these need windbg).

Re:Duh

[ihtoit]

I still get 0x124 bugchecks - but that's because I run the snot out of my hardware so it likes to overheat.

Re:Duh

[ihtoit]

if I had an SSD instead of a spinny thing, I'd be running all swap ops on an SD card. They're dirt cheap (8GB for what, three quid these days?) and swapping out a dead one is a simple case of press until it clicks and wait until the spring ejects it. Hell, you don't even have to power down to do it.

Re:Duh

[CronoCloud]

It's nice that you can replace the default Start menu, but in Linux if I don't like the defaults of KDE or Gnome, getting them how I want won't require a third party download.

Re:Duh

[spacepimp]

You know they fired Sinofsky because of Windows 8... Is that because consumers are all petulant assholes? Perhaps if MS wasn't so desperate to try and force their way into the tablet space, Windows wouldn't feel like a bipolar experience. Win 8/8.1 reeks of desperation. I give credit to MS for trying to push the OS forward, but I do not agree that the way forward was forcing in a touch interface to get to the desktop application which is anything but touch friendly.

Re:Duh

[Anonymous Coward]

You can add a swapfile instead of a swap partition in Linux, it will have a little more overhead.

You create a zeroed file using DD, use mkswap to ... make it swap? And then swapon /thenewswapfile

Re:Duh

[goarilla]

Fucknose why anyone would actually want that though.

I believe it was Alan Cox that uses a high swappiness (99) to keep interactive programs always fed with real memory.
The idea is that the programs with low cputime (deamons, ...) gets paged out before the interactive programs.

Re:Duh

[LVSlushdat]

Since I retired, I use Linux almost exclusively, but since I get bugged by friends/family/neighbors to fix their Windows booboos, I figured it was time to see for myself what abortion MS had wrought with Windows 8/8.1. Since I'd done some work for a neighbor who was short on cash, but had a spare boxed copy of Win8.1, I took the copy of Windows, kind of wondering what I'd ever do with it.. Decided to take a spare drive and install it on my Dell laptop. The install went fine, and, sure enough, this sure the hell doesn't look like XP (the last MS OS I *really* dealt with).. Installed a bunch of stuff with Ninite, then came the 100+ updates, and a reboot.. After the reboot, and login to the desktop, first thing out of the machine was kaboom, a nice lite blue screen telling me there was a problem and the machine had stopped.. A "VIDEO_TDR_FAILURE" in part of the nvidia driver that apparently had been applied in the mass of updates.. A bit of googly-ness showed that this was a common thing with the crap video drivers that Windows update carries.. So a visit to safe mode (no longer as easy to get to as XP's safe mode was) and an install of the latest driver from Nvidia itself.. Now the machine works nicely, with classicshell and some other tweaks that "fix" the idiocy that is Windows 8/8.1... But as for me, after this little trip into MS-land, I'm back to Linux....

Re:Duh

[goarilla]

Apparently It could have been Andrew Morton (https://lkml.org/lkml/2004/4/28/263).

Re:Duh

[goarilla]

I don't think that there's any evidence that the linux swapfile performs better - and in any case why would it being unfragmented be an advantage? Memory access is random, and so swapfile access is random, and so why does having it non-contiguous cause an issue?

That's exactly why you want a swap partition at the start of the drive where seek time is as low as possible.
You don't want a bad situation (swapping) to become even worse (swapping and seeking all around the drive).

Re:Duh

[operagost]

You are correct and the GP poster is wrong. NT 4 was quite stable as long as you set it up correctly (including using only the VGA driver, or maybe a very minimal ATI driver with no crapware (it seems like every server had an ATI Rage card back then) and left it the hell alone except for security patches.

Re:Duh

[Demonoid-Penguin]

Apparently It could have been Andrew Morton (https://lkml.org/lkml/2004/4/28/263).

That'd make sense - then. Even running a full blown desktop and a large compile I rarely see swap being touched.

On an old Thinkpad I have swappiness pushed to 80 or sound and video suffers, same era, limited RAM by modern standards. From dim memories I think I had to do a bit of niceness tweaking as well - price you pay for 512MB. I'd forgotten I even had that laptop till you posted that. (still works fine - updating to Jessie now).

Re:Duh

[MobileTatsu-NJG]

I'll concede to that. I've always used Nvidia for exactly that reason.

Re:Duh

[Anonymous Coward]

You know you can also use swapfile on linux? Partition is just the preferred way becouse performance

Re:Duh

[jimbo]

Mhmm, those assumptions are clouding the issue: you can easily put the Windows swapfile on any partition of your choice and have swapfiles on multiple partitions if that tickles your fancy. You can created a partition just for the swapfile if you want and years ago it was popular among enthusiasts getting a fast dedicated HDD just for the windows swapfile - before we all started putting gigantic amounts of RAM in our boxes.

But whatever difference the two approaches have between them in performance it's probably negligible compared to the penalty of using swap in the first place, in many cases anyway if not all.

Re:Duh

[cbhacking]

Definitely a whoosh, although there's actually some perks to the Windows NT kernel vs. the Linux one.

In any case, I use Windows significantly more than Linux (though I use both regularly), and the Linux I use is usually (though not always) in a VM and thus it doesn't have to deal with really weird hardware. Nonetheless, I get about the same number of kernel panics in each OS (1-3 per year, across three different regularly-used machines and various client loaners).

Re:Duh

[Grishnakh]

I don't know about other KDE users, but I've never "hacked around with settings all day"; I've spent maybe 10-20 minutes hacking around with different settings, trying things out, and after that I've just left them that way for ages, because at that point I'm happy with them.

So instead of spending less than a half hour to play around with settings in KDE, you want me to write my own shell program so I can use Windows??? Are you fucking nuts?

I don't have to do any programming to use KDE and change it around to my liking. This is supposed to be a plus for Windows?

You sound like all the idiotic Gnome sycophants: "if you don't like the default Gnome settings, just write an extension! And when they break the API in the next release, just recode your extension for it!"

Re:Duh

[Anonymous Coward]

Metro is a pile of shit. Admit it, who could you be kidding here?

Re:Duh

[drinkypoo]

But whatever difference the two approaches have between them in performance it's probably negligible compared to the penalty of using swap in the first place, in many cases anyway if not all.

This is more or less what I mostly came to say — how's that for convoluted. But to wit: Who cares whose approach to swap is "better"? Swap is crap. Most of us don't need any. We've beat this horse well beyond death here on Slashdot repeatedly. RAM is stupid cheap now. There are a few things that people do with their PCs now that take more than 8GB of RAM, like high-res video streaming from the same PC on which they're gaming or doing live video stream manipulation or whatever it is they're doing, and no amount of swap will help you do those things. All swap does it make your system thrash before it crashes.

With that said, the way windows handles paging is crap. It only lets you make one swapfile per drive, and you can't swap directly to a partition; mkswap is a lot faster than mkfs, or even an ntfs quick format, let alone the real thing. If you want more paging on the same volume with linux, you just create another swapfile and swapon to it at a lower priority than your primary file. When you no longer need more paging, you swapoff the file and you can delete it. If you let Windows manage the length of the paging file, then if that ever actually happens, you just wind up with fragmentation and that will impact system performance while swapping, for real. It will also impact your ability to defrag, since the paging file can't be moved while it's in use. You have to remove it, reboot, defrag, enable it, and reboot, since pagedefrag doesn't work after Windows XP.

TL;DR: the best way to manage paging on Windows is to disable it for all volumes. The best way to manage paging on Linux is to not create a swap partition or file. It's better to crash sooner and reboot than to crash later and reboot... later.

Re:Duh

[im_thatoneguy]

It's nice that you can replace the default Start menu, but in Linux if I don't like the defaults of KDE or Gnome, getting them how I want won't require a third party download.

Ummm... KDE and Gnome *are* third party downloads.

Re:Duh

[drinkypoo]

OS X users the same underlying functionality from a UNIX-like VM subsystem, but has a dameon that monitors the amount of used swap space and creates new swap files when they're required. This gives you the flexibility of the Windows model, without the complexity in kernel space.

If that daemon isn't a script, WTF. Because you could definitely do that with a very small shell script.

Re:Duh

[chipschap]

"you couldn't be bothered to spend 5 minutes to read or watch a Windows 8.1 tutorial to learn about all the nice features"

Um, well, the first "nice feature" was Win 8.1 refusing to connect to my home network. I dug into things and found that the as-delivered default was to connect to 802.11n ONLY. My wireless router does only 802.11g .... I was out of luck until I changed the default. Now, how many average users would have figured that out and fixed it? And you talk about "nice" features? By the way, I tried the network troubleshooter, and it failed to run because ... it couldn't connect to the network to fetch the help screens. Another "nice" feature, maybe?

In the end the huge shift in the UI paradigm, the bad reviews I've read, the bad experiences I've heard about --- coupled with my own initial frustration --- made me quickly decide it just wasn't worth it. Maybe I'll try Win 10 when it comes out, if it offers me any advantages in actually getting stuff done. That's why I have computers: to get stuff done, not to look at random advertising or have CNN headlines in my face, or to check the weather here in Hawai`i, when I know it's nice out without having some colored tile tell me so.

Re:Duh

[lucm]

Yes there is some serious incoherence in Microsoft's strategy, always been like that. Just look at mobile - on one hand they buy a handset manufacturer, on the other they fail to improve their suboptimal and antiquated mobile SDK. Or on Windows 8 they push the tablet paradigm, but they lock down their app store so badly that most developers give up on making Metro apps.

From what I've seen, with Windows 10 most of the metro fluff will be gone. Doesn't mean that the whole 8.0/8.1 thing was garbage; there were a lot of improvements besides the metro snafu.

Re:Duh

[TheRaven64]

Really? How does a shell script get notifications from the kernel that the swap space is almost full? Or are you suggesting that they should write a program that gets the notify(3) events, but then replace the three lines of C required to create a file and add it as swap with a shell script?

Re:Duh

[drinkypoo]

How does a shell script get notifications from the kernel that the swap space is almost full?

How does the daemon do it?

Or are you suggesting that they should write a program that gets the notify(3) events

Wait, you mean OSX doesn't have simple filesystem-based access to that, like /sys and /proc on real systems?

Re:Duh

[CronoCloud]

Only if you use something like Slack or Gentoo. Install something like Fedora and Gnome is installed by default.

I personally run XFCE and getting the interface the way I like it (panel on the bottom with application menu, quick launch items, window buttons, notification area and clock) doesn't require a third party download.

Re:Duh

[TheRaven64]

From this, I honestly can't tell if you're trolling or an idiot. Read the notify(3) man page and if you have non-trolling questions then let me know.

Re:Duh

[drinkypoo]

My point was that a very small shell script can detect memory conditions on most Unix systems through typical file access mechanisms. Apple will doubtless have overcomplicated this. I wonder if they have also created dependencies on that daemon. Does it do anything else?

Re:Duh

[JWSmythe]

You know, you can just put Classic Shell on, and it looks and behaves almost identically to Win7. The only significant change I see in 8.1 once you get out of Metro is the task manager is a bit better.

I know that making one change is a significant hardship to most people. It's probably not as hard as explaining why they can't run Windows programs. Or at least trying to explain why particular programs don't work under Wine.

Re:Duh

[brantondaveperson]

I wonder about that though.

If you're swapping, there's a very good chance that you're about to hit some other area of the disk in very short order. Given that the subsequent disk access is certainly guaranteed to be random, the optimum position for the swap partition is probably in the centre of the disk. Unless the heads changing direction is expensive (I've no idea), in which case the position of the swap partition may have no impact in the general case at all.

Re:Duh

[Evan Langlois]

LVM would have been best, but since its too late for that, you can add more swap using a file. Using a file is not best because you need to go through the filesystem layer and then the device layer, rather than just ignoring the filesystem. It has nothing to do with contiguous files. You can definately swap to a file as easily as as a device, and you can tell your system to swap to your existing swap device first before using the file (for speed). If you had two swap partitions on different physical drives, you would give them the same priority number so that Linux writes to them in parallel (like RAID). See man pages for swapon and fstab.

Re:Duh

[goarilla]

Mmm good thinking but you're assuming the swap partition is on the same drive as the system.
While this is assumption is right most of the time, so is it that the system partitions (/, /usr, ...), from which you'll be fetching that "new memory hungry application", also lie at the start of the drive.

Re:Duh

[Anonymous Coward]

Linux is nothing but 3rd party. If you can construct a Linux desktop environment with tools from less than a dozen projects your doing something very weird.

Re:Duh

[TheRaven64]

How does a shell script monitor for something crossing a threshold? You could have a file somewhere in /proc (or whatever) where blocking reads wait until it has crossed the threshold and then sends the notification. Only now you have to make sure that the priority of this shell script is high enough that it will always be scheduled to run before anything else that might allocate memory (and you need to ensure that your shell script interpreter won't allocate memory while running the parts of the script that are needed to create the file - trivial in a C program but basically impossible in shell).

Your in-kernel support now has to be a special-case file that supports both edge-triggered and level-triggered events, because the GUI task that prompts you to kill processes or resume processes once you've handled the OOM condition (Linux just kills a random process here, putting both policy and mechanism in the kernel and picking a particularly bad policy).

That's a lot of special-case logic in the kernel for one file. Or you can use a generic notification framework, such as notify(3), which allows you to get events via signals (if you want them asynchronously), via blocking kevent calls (if you want to integrate them with a generic event loop), or via a shared memory segment (if you want to poll for them) and also supports priority propagation, so the code in the VM layer is a single call saying 'deliver this event, it's really urgent'. The same logic is also used to send notifications at other watermarks, so (for example) NSCache instances can delete objects when memory starts to be a bit constrained.

Your solution is less flexible, doesn't really simplify the userspace portion (something that waits for a notify event and registers new swap sources and is carefully designed not to allocate memory is a dozen lines of C. It's a bit more if it wants to walk the process table and send SIGSTOP to processes that are using too much and send on another notification for the user, but only 30-40), and significantly complicates the kernel code. In other words, it sounds exactly like the solution I'd expect to see in Linux.

Re:Duh

[Tower]

In any case, can you make Linux use a swapfile permanently?

Yes, you just add a line to fstab to make it permanent. There's a brazillian guides on Google that will show you how.

But I don't read Portuguese, you insensitive clod!

Re:Duh

[Anonymous Coward]

How about OS/2 on a PS/2 - 1/2 of an operating system on 1/2 of a computer! :-)

Re:Duh

[Anonymous Coward]

I always wondered why they didn't call the TI 99/4 the TI 24.75. :)

I kinda miss smit from AIX

[leonbev]

I always wished that someone would make a smit like AIX control panel menu system for the Linux command line. If you couldn't remember the exact syntax of a command, it often came in handy for quickly getting the task done (without having to Google command examples) and then get an example of the syntax for next time.

Re:I kinda miss smit from AIX

[CanadianMacFan]

The first time I used AIX the administrator had installed it without the man pages! That was painful. How do you install a UNIX system without man pages? Mind you the other developers on my team wanted to run a shell script, as a child process, to set up all of the environment variables and they wouldn't believe me when I told them it wouldn't work. I didn't stay there long.

Re:I kinda miss smit from AIX

[lucm]

For most commands on AIX, ike wget, they managed to remove the useful options. And they still force you to have username of 8 characters max.

If one day there's an International Court of Justice for computers, AIX will be among the first to be on trial.

Re:I kinda miss smit from AIX

[mlts]

Closest to this on Linux is the SuSE tool, YaST. It is pretty decent, but only found on that distro.

Re:I kinda miss smit from AIX

[greenfruitsalad]

you've reminded me of my favourite MOTDs from fortunes package:

"Whip me. Beat me. Make me maintain AIX."
(By Stephan Zielinski)

"Problem solving under linux has never been the circus that it is under AIX."
(By Pete Ehlke in comp.unix.aix)

Re:I kinda miss smit from AIX

[Blade]

Except AIX has allowed up to 256 character usernames since 5.3, released in 2004.

Re:I kinda miss smit from AIX

[Blade]

Any OS I built would include AIX's LVM. Rock solid, conceptually simple for the most part, totally embedded into the storage management, and just works.

every scandal is a "_____-gate"

[turkeydance]

and all new combinations are "Franken-____". how about a Franken-gate?

Re:every scandal is a "_____-gate"

[oneiros27]

We're saving that for when the junior senator from Minnesota does something exceptionally stupid.

Re:every scandal is a "_____-gate"

[arth1]

and all new combinations are "Franken-____". how about a Franken-gate?

Better than a Waterstein, I suppose.

Re:every scandal is a "_____-gate"

[ChrisMaple]

Because the great disaster Al Franken is already in use.

100% genera

[Anonymous Coward]

(filler (filler (filler ()))

genera pretty much was the best thing ever

SOM and WPS

[martiniturbide]

Maybe some people do not appreciate it, but I think that having SOM (System Object Model) and WorkplaceShell (built over SOM) gave some functionality that was never exploited properly. So, building a Frankenstein I will put over the GUI, SOM (maybe the open NOM and somFree) and the Desktop elements will be built over it. (like Workplace Shell). Check out this OS/2 reference graphic: http://www.edm2.com/index.php/...

Re:SOM and WPS

[tlolczyk]

SOM would be cool. But hey even COM would be better then dbus.

I want...

[Anonymous Coward]

I want an OS with the hardware and software support of Windows, dmg style installer from Mac OS X, ZFS from OpenSolaris/FreeBSD, and a great command line experience from pretty much any non-Windows OS.

Re:I want...

[slaker]

For what it's worth, Powershell on Windows is actually pretty nice.

Re:I want...

[BevanFindlay]

You do realise that most of those wishes are granted with any modern Linux install? Hardware support has gotten a lot better (mostly it's just "install and go" now), software support is either (a) native versions of the stuff you want, or (b) installable using WINE (not everything works well with WINE, but it also is much better than it used to be). Installing software on Linux is in my opinion easier than most OSes, as long as it's in the main catalogues: just go to your software manager, do a search, click install. Even for more obscure stuff, it's maybe just adding a repository, which is a simple "Google for it, then copy and paste a line or two of text". Apparently, Linux also has native ZFS support.

Or am I missing something here?

Re:I want...

[ChrisMaple]

Last time I looked at WINE, it couldn't deal with devices on USB ports, and the developers explicitly stated they had no intention of ever making USB ports work.

Re:I want...

[dave420]

Linux hardware support has got a lot better, as has software, but to pretend for one instant it's comparable with Windows is lying to yourself.

Re:I want...

[Anonymous Coward]

Place the blame where it belongs: a proprietary OS.

Re:I want...

[jones_supa]

If Wine was commercial software, the customer demand and highly-motivated crack team of paid developers would have resulted the software to have premium USB support 10 years ago.

Re:I want...

[Anonymous Coward]

You are missing something. It's called perspective and the ability to think outside of the box you are currently in.

Re:I want...

[Evan Langlois]

Yes, so true. Windows is a toy used to run games and spread viruses. Linux is what you use to get work done, from the million Android phones to the worlds largest super-computers, with desktops, tablets, and even your TV thrown in to the middle. Just plugging in a USB device into Windows is like pulling teeth - you have to let it search for drivers that it never finds and then finding a damn app that doesn't download a virus along with it is nearly impossible. And the fonts look like something from the 90s. Its a joke. Under Linux .. plug in cell phone, MTP file system shows up (and no stupid green bar locking up the desktop when I open the device like under Windows). A tap on the phone turns on USB networking and Linux sets up the network device and gives me a notification that it succeeded while removing the MTP icon (the phone won't do both at once). Tap the USB tether icon on the phone again (which is Android, also Linux), and the USB network is taken down and the MTP file system is back. Don't try it under Windows. I end up screaming at it while it looks for drivers or gives some stupid never-ending progress bar. Hardware support under Linux beat Windows hands down.

Re:I want...

[BevanFindlay]

No, not really. I spend more time on Windows than on Linux. In fact, until a few months ago, I hadn't touched a Linux install for a few years, however I was quite surprised at how slick it has all got now when I did get back to installing Mint on a machine I had. I was just pointing out the fact that really the "Which OS are you using?" question is starting to go the way of "Which web browser was this site designed for?" - i.e. it kind of doesn't matter any more. No, Linux' device support isn't perfect, and no, support in WINE for other software isn't perfect, but it's all a lot better than it used to be, and covers most of the requests that the post I replied to was asking for.

Genode and Reactos

[Anonymous Coward]

I would take the simplicity of Genode, the UI from KDE, and the driver layer from ReactOS. Added security with your choice of micro kernel, the beauty of KDE for the non "power users", and the ability to use the hardware you paid for.

...game

[Anonymous Coward]

The parts that let me play games.
Disregard everything else.
I would love for a lightweight OS that only has the dependance for gaming components like DirectX, OpenGL, etc.. While at the same time being x86/x64 compliant so I can install the applications I want on it.
Windows as it is, is so filled up with junk that it could do with getting the fat trimmed out.

Re:...game

[Anonymous Coward]

So... You want a 64-bit clean DOS with OpenGL?

Re:...game

[WilCompute]

That sounds better than you think. I wonder has anyone thought to port Vulkan to 64bit freedos.

Re:...game

[Dr_Barnowl]

Yeah, what I've wanted for a long while is basically "X Box OS for Desktops" (but hopefully without the advertising bloatware...)

Re:...game

[jones_supa]

SteamOS is like that.

Re: ...game

[Anonymous Coward]

That would be great! Not that I'm a gamer, but I do some extremely processor intensive simulations (think days on end with the CPU maxed out) and would love to see how they fly when you get all the cruft out of the way.

In fact, forget OpenGL and just give me an absolutely minimal, bare bones os.

Rather Than in more out

[rtb61]

Custom OS is not about what is in it but all about what is left out. Custom OS for appliances that only has in it what is actually necessary for that appliance, maybe just maybe incorporate application into the OS rather than a separate post boot load.

I think future trend will be a shift from more flexible universal operating systems to more modular, take every out that is not necessary for this particular appliance operating system, this to simplify security and even application level features become modules added into the operating system, so one quick boot to full functionality. So a much more modular operating system.

Re:Rather Than in more out

[i.r.id10t]

And for an single tasking appliance, that is fine. Remember the i-opener from Circuit City? Ran a very compact and relatively fast (for the time) QNX with just a crappy browser like application on top of it. Of course, it got hacked around, spawning the "i opened it" small computer/display setups. Quite a few /. articles/postings on this.

Re:Rather Than in more out

[PopeRatzo]

Custom OS is not about what is in it but all about what is left out. Custom OS for appliances that only has in it what is actually necessary for that appliance

Your mouth to Microsoft and Apple's ears. I want an OS that does nothing but run my programs and stay out of my way. I can get my own browser thanks.

If y'all could just get Linux to run current AAA games, and some professional music & audio software, I'd never spend another nickel with Microsoft or Apple.

Re:Rather Than in more out

[Anonymous Coward]

A big reason for this difficulty is because Windows and Apple provide APIs that are used by these programs, which is why porting them takes work and isn't done.

Re:Rather Than in more out

[dave420]

Which changes absolutely nothing. The software is still not available on Linux. Excuses are nice an' all, but change squat.

Re:Rather Than in more out

[Anonymous Coward]

Don't know what an "AAA" game is - but you can certainly have pro music & audio sw on linux. Not the SAME as on windows, but pro.

Re:Rather Than in more out

[PopeRatzo]

but you can certainly have pro music & audio sw on linux.

You can have music and audio, but not professional. The quality of the programs is just not there yet, sorry.

Re:Rather Than in more out

[Lesrahpem]

I think future trend will be a shift from more flexible universal operating systems to more modular, take every out that is not necessary for this particular appliance operating system, this to simplify security and even application level features become modules added into the operating system, so one quick boot to full functionality. So a much more modular operating system.

A little OT, but I think modifying Gentoo with a new build system to do the above would be a fun project.

Re:Rather Than in more out

[rapu]

Although typically, you wouldn't make an OS run the software, you'd make the software run on the OS.

Re:Rather Than in more out

[Evan Langlois]

I don't agree. http://kxstudio.sourceforge.ne...

Re:Rather Than in more out

[PopeRatzo]

That's just packaging the same old wine in a new bottle. Linux is still way behind in hardware support, has only the most limited VST support and the DAWs are still nowhere near ready for prime time professional production.

It's getting there, but there's still a ways to go.

Every year for the past 10 at least, I've made a run at professional audio production on Linux, and every year I'm disappointed. It's not like Linux is not useful in the studio and post-production. In fact, after Cockos Reaper came out with ReaMote, allowing for off-loading of samples and streams and effects, there has always been at least one Linux box in my production chain. But as main production machine, the applications just are not there yet. And as long as kxstudio relies on Jack, it won't be there.

Re:Rather Than in more out

[Evan Langlois]

1 VST is commercial. You can't blame Linux because some other company isn't making their software available for it. 2 You can't ask for pro-audio and then say that you don't want jack. Jack is not only your best option on Linux, but its good enough for Apple! They use jack, too. You are comparing with Windows, and thats the worst option. Windows latency is way higher than Apple or Windows. Hardware support on Linux is great. I plug it in, it works. Windows on the other hand is a driver nightmare! You try downloading drivers and get viruses instead. I shouldn't even have to download the crap - it should find its own drivers, but that has NEVER worked for me. Sabayon took about 30 mins to install on my laptop and autodetects everything. Windows was an all day thing having to transfer drivers via USB off my phone because Windows didn't detect the network card (not wired or wireless). Still don't have the accelerometer driver under Windows, but its detected fine on Linux.

Re:Rather Than in more out

[PopeRatzo]

1 VST is commercial. You can't blame Linux because some other company isn't making their software available for it

VST is professional. I can't help it that a musical instrument or high-end compressor/limiter costs money. Until there are such tools available for Linux, at professional quality, it's just not going to be a platform that a music/audio professional is going to use (with very rare exception).

Hardware support on Linux is great. I plug it in, it works.

You're still talking about consumer gear. You cannot make your Universal Audio or RME interface work with Linux.

Jack is not only your best option on Linux, but its good enough for Apple! They use jack, too.

You find me one single audio professional that uses Jack ported to Apple. Seriously. Yes, it's been done, but nobody's using it professionally.

And latency on Windows is not "way higher" than Apple.

Re:Rather Than in more out

[PopeRatzo]

1 VST is commercial. You can't blame Linux because some other company isn't making their software available for it

And by the way, I'm not "blaming Linux" for anything. I have high hopes for it to become a pro-level music/audio platform. I've contributed to the Cockos Reaper for Linux project and annually try out the Linux audio tools to see what kind of progress has been made. It's just not there yet.

Try and make an OS that viruses couldn't target.

[Anonymous Coward]

I'd start with a Linux version where I figure there are no known major security holes.

I'd add the ability to run Windows binaries in emulators, but they can't access other programs than themselves. If that was a problem, add a phantom disk image so it could see other files that you place in the phantom disk image. Imagine each Windows emulated program saw their own personal c:/ , and it and you can populate it with files.

I'd make sure nothing outside the OS can run in the background process, unless a security approval by the user, and an inactive window can't can't read keystrokes.

If you wanted a program to start up on bootup, you'd need to boot to a special safe mode to allow it. And this is where drivers would be installed too. Drivers would be restricted to their own filesystem and can't write over other drivers.

Allow it to run a browser and based apps too, probably in the emulator Windows.

I figure if the software you download can't get out of the Windows emulator or its own personal filesystem, it can't mess with your OS or the rest of your filesystem. If it can't record your keystrokes unless you have the window actively open, a keylogger can't get you either. The problem is that we probably don't have perfect Windows emulation. Another problem is you have to be able to trust your drivers or that is a possible vector to an attack.

JCL

[Skewray]

As long as I can use IBM's JCL: https://en.wikipedia.org/wiki/... to run my jobs, I know I can be truly productive.

Re:JCL

[plopez]

JCL.... the horror...... the horror.....

Re: JCL

[Anonymous Coward]

JCL is awesome! As long as you never make mistakes...

Re: JCL

[Anonymous Coward]

Die. Die. Die.

Just in case you do not understand: ARRRRGHGHTHFUFJFJFJFJFHHGGGHHFHHHHH!

No. JCL.

Friends don't let friends take jobs involving JCL. They kidnap them and get them married to the nearest available fertile female knowing that their future will be far happier.

Bastard.

Re: JCL

[RabidReindeer]

JCL is awesome! As long as you never make mistakes...

Ah yes. "If return code from STEP2 is NOT LESS THAN 4, Then DO NOT execute STEP5".

Re: JCL

[Anonymous Coward]

My mum used to wrire j l you insensitive clod.

Easy

[rev0lt]

FreeBSD kernel, Solaris networking/clustering capabilities, and a sort-of Windows UI (Imagine Windows 7, not the metro interface)

Apps, apps, and only apps!

[Anonymous Coward]

App an app using apps, apps, and more apps! No Luddite Operating Systems belong in Modern systems, just apps!

Apps!

From a user standpoint

[Anonymous Coward]

I'll give you a viewpoint from a user perspective for what would make an ultimate frankenos:

Speed: Atari TOS or BeOS
Multitasking ability: Amiga or BeOS
Stability: Insert reliable *nix variety of your choice
UI: Mac OS X Snow Leopard

I can't imagine how great computing would be if we could have kept the best of each instead of ending up with the "juggernauts of compromise" we widely use today.

Re:From a user standpoint

[Anonymous Coward]

OS X Snow Leopard is super fucking overrated. The "grandpa" version of OS X.

I'd rather use Yosemite, the latest release (10.10.4, with discoveryd gone), any day of the week.

Re:From a user standpoint

[TheRaven64]

Try this. Launch the same applications on Yosemite and Snow Leopard and see how quickly to can tell which one is the currently active one. After SL, they dramatically reduced the visual clues (the big shadows on the active window that made it stand out were 'ugly') and they've reduced more each release. After SL, the instances where I typed things into the wrong window jumped up for me. It's a shame, because Apple used to be the company that measured this stuff...

Re:From a user standpoint

[damnbunni]

I updated my Mac mini from Snow Leopard to Yosemite.

So far as I can tell, the only thing better about Yosemite is that you can turn off the transparency effects.

That's it. If it wasn't for software that needed 10.7 or newer to run, I'd fuck right off back to 10.6.8.

Re:From a user standpoint

[Phreakiture]

Additional things to bring from Amiga:

• From the CLI, a directory appears to be just another executable object. Name it on the command line and you are there. Put another way, the "cd" command is just white noise.
• Device names, volume names and aliases are interchangeable. Program or script needs a particular volume of removable media? Simply reference it by name (e.g. foo:). A dialogue box will pop up asking you to put that volume into any drive and click OK. You never need to click OK, because it will detect the volume insertion and carry on. Did you copy that data to the hard drive? Just define an alias so that, for instance, foo: now points to system:volumes/foo.

That said, don't look to Amiga as a model of stability . . . it wasn't. All it took to bring one to a red guru meditation was to attempt to read past EOF.

Re:From a user standpoint

[Anonymous Coward]

Yosemite spares me the effort of having to click "save" every time I make changes to a document or picture.

Unfortunately, it replaces it with the effort having to click "restore" every time it saves some change I wasn't aware I made. :-P

Do not need the 'franken' part.

[Demena]

Plan 9 works just fine.

At true "FrankenOS" would have...

[mykepredko]

I would expect that it would just have a big screen, continually scrolling messages like:
- Rush Limbaugh is the devil incarnate
- The Republicans are all owned by the Koch brothers
- The Bushes practice devil worship
- Donald Trump is ashamed of his Mexican heritage

Or, do you mean "FrankenOS" in terms of "Frankenstein" and not "Al Franken"?

Re:At true "FrankenOS" would have...

[Anonymous Coward]

> - The Republicans are all owned by the Koch brothers

That's systemd...

Re:At true "FrankenOS" would have...

[Anonymous Coward]

mykepredko actually thinks there's any real difference between democrats and republicans? it's hilarious to find these dinosaurs

The Big Three

[YukariHirai]

Currently, what I'd really like is something that's mostly a typical GNU/Linux system, with OS X's GUI and reliable support for Windows and Mac applications.

Re:The Big Three

[Anonymous Coward]

Isn't that just OS X with specific software (e.g. VMWare Fusion ), or is there a GNU/Linux vs FreeBSD issue in the under the hood userspace ?

Re:The Big Three

[lucm]

OS X's GUI

That one is easy. Just find a 15 years old version of KDE.

Re:The Big Three

[WilCompute]

Well, I would imagine he wants the GNU stuff to be able to be swapped out for newer versions without breaking OS X at ever OS update. Or maybe he wants the GNU compiler to be included again. Or maybe he wants the GNU tools to be first class citizens, instead of slowly being fazed out for applications written by Apple.

Then again, maybe he wants the Linux kernel, with the GNU license, instead of the Apple License the Darwin (which is not FreeBSD, though it has some shared history). Or maybe he wants a kernel that wasn't forked from BSD at the time the were trying to create a micro kernel by tacking things on. Or maybe I am just guessing, and I don't know what he wants.

Re:The Big Three

[morgauxo]

Getting that OSX GUI exerpience without OSX is actually pretty easy!

First... chop off your dominant hand.
Also, remove all but one finger from your remaining hand.
Put a patch over your dominant eye.

Finally, to complete that Apple feeling get a friend to kick you in the balls every 5 minutes.

Voila! It's just like using the OSX UI!

Re:The Big Three

[Anonymous Coward]

Sounds like a Mac. Just get a Mac.

Re:The Big Three

[IMightB]

It still sounds better than Metro...

Re:The Big Three

[LWATCDR]

You have hit on a truth that is old as Byte Magazine.
OS/s for the desktop never really mattered it is all about applications. You buy a computer to run applications you do not buy a computer to run an OS.
I would love a desktop with VMS running and OS/X UI and possibly some OS/2 Workplace shell as well but without the apps it would be useless.
It is kind of like what happened to the Atari ST and Commodore Amiga. They both had a better OS the MS-DOS and they both had better hardware at a cheaper price than you could buy to run MS-DOS.
The lack of app combined with the lack of coverage in the computer magazines is what killed them. You can not blame the magazines too much since Atari and Commodore could only buy so many ads compared to the all the adds that the clone makers could buy.
Right now I run Windows 7 at work but have VMs for Windows 8 and Linux for projects I work on.
ChomeOS has a chance now that so many people use their computers for the internet and not a lot else. I do not know the last time I bought a piece of software that was not a game for a PC.

Re:The Big Three

[morgauxo]

You just might have a point there!

Re:The Big Three

[YukariHirai]

Honestly, sounds more like the Windows experience to me.

Re:The Big Three

[Evan Langlois]

Darwin shares no history with FreeBSD nor was it a fork. It is a modification to an MIT microkernel called Mach (same kernel GNU Hurd was slated to use). Normally, the unix compatible services were to be excluded from the kernel. Mach only does resource management and IPC. For performance reasons, they included a BSD-compatible "server" into the kernel, making it a monolithic kernel. This design was then included as the kernel used by NeXT. OSX is just a bloated, prettier version of NeXT Step. Its NOT BSD!

Re:The Big Three

[Anonymous Coward]

Darwin is every bit as much a BSD variant as FreeBSD, NetBSD or OpenBSD is. It can trace its roots in both design and code directly back to 4.3BSD. It's basically FreeBSD with a different kernel and driver model.

You're aware that the Mach microkernel was originally designed as a drop-in replacement for the BSD kernel? They abstracted more functions into the microkernel as the project went on. MachOS was always intended to be a BSD system, with the option of running other environments on Mach in the future.

NeXT was BSD before the Great Divide from AT&T code. Jobs paid a license fee to AT&T to use Unix.

Take a look at the Unix family tree and note where Mach, NeXTSTEP, Darwin and Mac OS X reside.

Mac OS X is Unix, specifically a variant of BSD Unix. Not exactly sure why this is a problem for you, unless you helped build BSD and NeXT killed your father?

Re:Try and make an OS that viruses couldn't target

[khellendros1984]

The problem is that we probably don't have perfect Windows emulation.

Of course we do; the emulator just emulates PC hardware, and Windows runs within that environment. Did you mean that we don't have a perfect open implementation of every Windows API?

Re:Try and make an OS that viruses couldn't target

[viperidaenz]

Can you explain how a keylogger in Windows gets access to user input in another application without a security approval by the user when it was installed?

Hint: It needs to modify kernel structures on every reboot or install a driver to do so.

Re:Told my nephew @ Apple this... apk

[viperidaenz]

Sorry, your post has been blocked by a HOSTS file.

Is this a troll?

[mykepredko]

I'm asking because I can't think of a consistent set of features that I would want in "my" best OS. I don't think there's any hope in consensus among /.ers and I don't think that individuals could come up with a single system that encompasses the different uses they put a computer to.

I would want different things from the OS depending on the activity I'm doing on a computer at a given time:
- coding.
- business activities.
- playing games.
- playing media.

Maybe there are features that consistent between all of these activities, but I would think that for the most part different usages would dictate different features and functionality.

Re:Is this a troll?

[WilCompute]

Well, while we are getting rid of the old OS, so all the applications need to be recompiled anyway, lets switch to the Mill Computing Platform. Why? Number one feature for an OS, there is no such thing as a ring zero. All applications, including the kernel, run at the same pace.

Now, couple that with all the benefits of the Mill Architecture.

* Next, lets start with seL4 as the Kernel, as the only mathematically proven, verified bug free kernel*.
* Add Vulkan as the only Graphic API,
* Switch to ZFS for large clusters of Hard disks,
* Switch to an SSD optimized Filesystem for appropriate hardware,
* Add different flavors for the OS. seL4 has this idea that the OS interface is simply made up in how applications pass information. So, you can setup a Windows passing scheme, a Genode passing scheme, a gaming specific passing scheme, an OS X passing scheme and so on and so forth.
You can then have applications written directly to the metal, or locked to one particular scheme.

Yes, I know that it doesn't work quite like this, but we are dreaming, aren't we?

* Yes, this assumes no bugs in the verification and all that jazz.

Win X

[Anonymous Coward]

The MS Windows interface on top of a Unix. Kind of like OSX.

MS-DOS's error message

[Anonymous Coward]

Abort, retry, fail, ignore?

Re:MS-DOS's error message

[ArcadeMan]

Keyboard error or no keyboard present. Press F1 to continue.

File versioning and backup flags

[Nkwe]

VMS supports very elegant file versioning, which I found a very useful feature. By default, every time you open a file for write (and you modify the file) you get a new version of the file (kind of like copy on write at the file level). When you list the files in a directory, you can see all the versions of the file with the version number being an actual part of the file name(file.text;1, file.text;2). On a per file basis you can set how many versions you want to keep around and the file system will automatically purge the old versions as new ones are created. When opening a file you can specify any version you want, or if you don't specify a version you just get the latest.

The file system also supports specific backup related attributes that integrate with the backup system. This lets you specify that a file should be excluded from backup and if I recall, tracks if the file has been modified since the last backup.

Re:File versioning and backup flags

[Anonymous Coward]

I concur.

The VMS versioning system was simple and elegant. If there was a problem, you could diff the versions. We had somebody who work late, put in a new version of a program that had problems. He would arrive late, so we could not check what he had done. We would revert to the previous good version. It was the best and simplest versioning system that I have ever come across.

Re:File versioning and backup flags

[tlolczyk]

The VMS versioning system was more of a kludge then something elegant.

For one thing, it versioned intermediate files too. Causing people to run "purge" (IIRC) all the time For another thing, if you wanted to keep an old version of a file, fine, but it might x.x;11 and the present version is x.x;40. It is far easier to find if you simple copy the file into another name before making changes. If you want to keep versions of several files, there exist several programs that do that. We call them version control systems.

Re:File versioning and backup flags

[Anonymous Coward]

VMS copied the feature from older systems. I remember it from the CDC machines in the late 60's (NOS/BE). As it is so often the case, "new" features are really a new implementation of old ideas (VMWare vs. hardware virtualisation in OS/VS on IBM mainframes, HTML vs. block mode terminals such as 3270 or even older PARS in the 50's and so on).

Re:File versioning and backup flags

[laird]

VMS versioned file system - YES! I don't know how many times that saved me.

Digital UNIX clustering was awesome. You could literally plug in more servers into a shared bus, and processes would automatically start on the new servers, with no reboot, reconfigure, etc.

Also BeOS' efficiency and boot times. You could cold boot BeOS faster than other OS' do a "resume from sleep". And you could decode and play multiple A/V streams at once in windows, while dragging them around, back when other OS's could barely decode a single stream to full screen (the easiest scenario).

Re:File versioning and backup flags

[Fweeky]

DragonFlyBSD's HAMMERFS does much of this - you can examine the version history of files and directories using hammer history and undo commands, and reference versions directly by appending @@ to filenames.

You can control how long history is preserved for and in what level of detail, as well as efficiently replicate it all across the network to remote filesystems (which can have their own, different rules). All this in addition to the more traditional named snapshots approach you're limited to with, e.g. ZFS.

https://www.dragonflybsd.org/d...

Re:File versioning and backup flags

[Fweeky]

Erm, "appending @@<version> to filenames".

Re:File versioning and backup flags

[Anonymous Coward]

Bah. From the perspective of 99% of the users, VMS versioning was simple, useful and automatic. Most VCSs are ugly firedemons from hell. You suggest trading a good system for a bad one. I suggest you re-read the OP.

Re:File versioning and backup flags

[Anonymous Coward]

The VMS versioning system was more of a kludge then something elegant.

For one thing, it versioned intermediate files too. Causing people to run "purge" (IIRC) all the time
For another thing, if you wanted to keep an old version of a file, fine, but it might x.x;11 and the present version is x.x;40.
It is far easier to find if you simple copy the file into another name before making changes. If you want to keep versions of several files, there exist several programs that do that. We call them version control systems.

Right. Because manual processing is so much more convenient than something automated? And it was easy for the administrator to run a purge program. I ran VMS based servers in a legal environment for 11 years and never lost a file because of this kind of capability.

Linux Mint + Windows Games & Photoshop

[Harlequin80]

That would pretty much get me right there.

Linux Mint & Cinnamon is by far the best UI out there at the moment. If it could flawlessly play windows games and things like photoshop I would be a happy man. All the other decisions feel about right. Ext4 for the filesystem, network manager is clean and works well, I love the file manager, bash shell and I'm pretty much done.

I'd leave ZFS for dedicated storage boxes.

Re:Linux Mint + Windows Games & Photoshop

[BevanFindlay]

Yeah, pretty much this. Also, I find that a much under-appreciated feature of Cinnamon is that it has a working "do not let an app steal focus" option; it is really quite amazing how much under most OSes, we get used to something being able to interrupt us mid-typing. Of course, I do sometimes have to get used to the fact that this can also mean that the app I wanted to start only appears on the taskbar and not on top of what I'm doing, but I'd far rather than that have something jump to the front two or three times while it opens, like I get on Windows.

Re:Linux Mint + Windows Games & Photoshop

[rev0lt]

I'm actually not a fan of Linux, but basically Mint is - by far - the most usable *nix desktop I've ever tried. If they could solve the out-of-the-box wtf-is-this problem with the fonts in Java (I know, its not really the project's fault, but the nineties called and they want their defaults back), I'd probably be running it right now. And I'm a BSD guy. On the other hand, what the f*ck happened with XDMCP? Everytime I tried it, its broken either by design (lxde login manager?) or just broken (gnome).

Re:Linux Mint + Windows Games & Photoshop

[Harlequin80]

Absolutely this. It causes me no end of frustration when Thunderbird steals focus because of it's dodgy calender implementation.

But I will now quite happily roll mint out to random relatives on a dual boot system and let them run with it. Most of the time they don't end up going back into windows because the most intensive thing they do is run Chrome.

Re:Linux Mint + Windows Games & Photoshop

[Harlequin80]

Where in Java are you talking about? I don't notice anything on my system...

As for XDMCP I don't think it is supported via MDM, the default login manager. However if you replace MDM with LightDM you will have working XDMCP support if you add the following to lightdm.conf

[XDMCPServer]
enabled=true
port=177

MDM dropped support for XDMCP back in 2013 and has no intentions of reimplementing it......

Re:Linux Mint + Windows Games & Photoshop

[Anonymous Coward]

Java applications never took off like they should have because of the write-once-look-like-shit-everywhere problem

Re:Linux Mint + Windows Games & Photoshop

[rev0lt]

Try installing NetBeans, it will default to some ugly-ass fonts.

Re:Linux Mint + Windows Games & Photoshop

[Anonymous Coward]

Gnome does this by default. You need to use an extension to let apps steal focus. You get a notification when the app appears instead of losing your focus.

Re:Linux Mint + Windows Games & Photoshop

[Anonymous Coward]

Use a different Java. My fonts on Java apps (like my Cisco ASDM) aren't as nice as my Linux desktop, but then again, Windows fonts are 100 times worse. I don't know how you can call Word WYSIWYG since the fonts are so damn horrible. I end up installing MacType just so I don't throw up.

And how is Gnome broken? I run Gnome 3.16 with GDM as the login screen and really like it. I've been using Gnome since it was first released way back when it was just a desktop based on the Gimp library. *ALL* my apps are Gnome apps (except ASDM).

From Puppy

[innerpeace]

Autologin + run as root. Two great features.

Mac

[Anonymous Coward]

MacOS X but with a MacOS9 style finder (as far as browsing, system folder, extensions etc).
Proper NTFS write support

That is easy.

[hairyfeet]

Win 7 GUI (with Win2K Pro GUI as classic mode) with WinFS** from Longhorn Alpha and WIMBoot from Win 8.

**-What is sad is that the WinFS demo was in 2003 and here we are in 2015 and we have YET to see anybody give us what they showed in that demo! Where is the file system that can just scan vids and pics and assign metadata like "girl in blue dress" or "white truck" and let me find any file by using human expressions without requiring the user to fill in the blanks?

Re:That is easy.

[BevanFindlay]

The sad truth is probably that the wonders of WinFS were actually nothing more than vapourware - something that looked like it might work on a very limited subset of possible inputs but which was found to be completely unworkable when faced with the real world of untidy data. Of course, I would hazard a guess that we are probably in a space now where image recognition and machine learning are at a point where that metadata-filling and searching is actually possible - it might not be all under the same patent portfolio though. Google is probably the closest, at a guess. It would require sending all your data (or hashes of it) to external servers for processing though, and that might be one of the limiting factors. Yes, it's the 21st Century, but not quite that much yet.

Re:That is easy.

[Anonymous Coward]

Here's the depressing story:

WinFS team was told that Longhorn (the version after XP) had to ship in 2005, so that's when WinFS had to be done. They weren't able to finish in time, instead finishing in 2006. Longhorn turned into Vista, which shipped in 2007. Because of Vista's schedule slip, we totally _could_ have had WinFS (if you were willing to take an 8-12% perf hit, too).

Re:That is easy.

[rev0lt]

I've taken Longhorn for a spin like everybody else. Most of it seemed to be done in managed code. I find almost no resemblance to a typical Vista system, it seems they rewrote a huge part of the operating system (again). Or they just glued some concepts together (in c#) to show some progress. Either way, WinFS always seemed vaporware.

Re:That is easy.

[Anonymous Coward]

> The sad truth is probably that the wonders of WinFS were actually nothing more than vapourware - something that looked like it might work on a very limited subset of possible inputs but which was found to be completely unworkable when faced with the real world of untidy data.

I was in grad school at the time.

IMHO, the problem with WinFS is that they were trying to replace the entire files system (NTFS) rather than just offering a relational service (SQL-lite).

Re:That is easy.

[Ralph Spoilsport]

Sure that's fine until you want to see how big your directories are - still epic fail on some thing Mac has done since forever. My hybrid dream OS: Linux kernal, MacOS UI modified by Win7.

Re:That is easy.

[Anonymous Coward]

The answer to your question is: a video game series called Halo.

Okay, that answer didn't sound very fair so far, so allow me to elaborate.
I haven't seen that WinFS demo, but I'll answer based on what I've read here, and what I've read elsewhere over the years.

After WinXP, Microsoft was going to release new technologies, called Longhorn.
Due to the substantial security problems with XP and Server 2003, resources were moved from Longhorn to Springfield, which was mostly a security update. The Springfield updates became released for free, and was called XP Service Pack 2.
Some of the newer technologies that were going to be in Longhorn got moved out of the OS updating project, so that Springfield could be reasonably released.
The WinFS was supposed to contain database functionality in the filesystem. The database and the filesystem were basically going to be combined somehow.
For whatever reason, perhaps due to technical feasibility or due to marketing forces, the decision was made to stop moving in that direction. The filesystem and database technologies got de-integrated.
Microsoft developed the database functionality further, and ended up releasing Microsoft SQL server. I'm sure that this database can keep track of things like fields.
As for being able to just find stuff on your hard drive, Microsoft seems to have released some technologies for that, too, called Windows Search or Windows Indexer. Notice that Start buttons now have a Search field, and can find files like that.
So, for your question (which I hereby summarize as "where is this technology?"), the answer is, "all over the place", including a commercial product called SQL server. The idea of having all of this bundled into one product seems to have been scrapped (probably over a decade ago). Things may merge and become integrated again, but doing so will probably be based on new, fresh efforts, unrelated to any of the ancient progress that people dabbled with so many years ago, because the working technology today probably looks quite different than the state that the technology was in back then.
Actually, the idea of trying to search for descriptions is something that I've experienced with some of Google's technologies, including the "OK Google" software. I understand Siri is similar in capabilities, and Amazon is interested in joining the fray. Microsoft decided they will release their own solution of being able to search for stuff based on that description. And the name of Microsoft's software that will be performing that feature is "Cortana", named after a character from Halo. If you're not a video gamer, you can look for it in upcoming Microsoft technologies, including Windows 10.

Re:That is easy.

[hairyfeet]

Well from what I read they had it up and running in alpha stage and that they basically rewrote a HUGE chunk of NTFS and combined it with MS-SQL. The problem? Forbes.

Ballmer kept promising that Longhorn would be just "whipped out" like they did with 95-98-98SE-ME but the actual engineers was trying to make real progress by coming up with a truly "next gen" OS. Ballmer saw it would take 5 years to get it out, got ridiculed in Forbes for being behind, and had them shitcan ALL the work they had done and just slap a new skin (And a bunch of crappy DRM, because Ballmer was ALL about DRM) onto Win2K3 and voila! Vista.

Of course we all know what a disaster THAT was, 3 months of fighting Vista was enough to convince me to stick with XP X64 (which was 2K3 X64 minus DRM and bullshit) until Win 7 X64 came out, but you'd think that somebody would have used such a great idea, but nope, instead we get more iCandy on OSX and Linux? More text editors and bad ripoffs of the Win/OSX GUI, in other words SSDD. Its a damned shame though as with today's multicores a FS that did like WinFS would not only be doable, I doubt the average user would even feel the 5-10% CPU on the initial run.

Re:That is easy.

[Anonymous Coward]

See BeOS and the Be File System .. 1996 ...

Re:Told my nephew @ Apple this... apk

[lucm]

Can you imagine being the nephew... Having this kind of technical advice on a regular basis, PLUS working at Apple. Se-weet!

Singularity

[WittyName]

Built on type safe language, so no buffer overflows. Layered on a type-safe assembly language. Immutable everything. Defined channels between components. No memory sharing anywhere.

Our recent article in Operating Systems Review, Singularity: Rethinking the Software Stack, is a concise introduction to the Singularity project. It summarizes research in the current Singularity releases and highlights ongoing Singularity research.
Overview

Singularity is a research project focused on the construction of dependable systems through innovation in the areas of systems, languages, and tools. We are building a research operating system prototype (called Singularity), extending programming languages, and developing new techniques and tools for specifying and verifying program behavior.

Advances in languages, compilers, and tools open the possibility of significantly improving software. For example, Singularity uses type-safe languages and an abstract instruction set to enable what we call Software Isolated Processes (SIPs). SIPs provide the strong isolation guarantees of OS processes (isolated object space, separate GCs, separate runtimes) without the overhead of hardware-enforced protection domains. In the current Singularity prototype SIPs are extremely cheap; they run in ring 0 in the kernel’s address space.

Singularity uses these advances to build more reliable systems and applications. For example, because SIPs are so cheap to create and enforce, Singularity runs each program, device driver, or system extension in its own SIP. SIPs are not allowed to share memory or modify their own code. As a result, we can make strong reliability guarantees about the code running in a SIP. We can verify much broader properties about a SIP at compile or install time than can be done for code running in traditional OS processes. Broader application of static verification is critical to predicting system behavior and providing users with strong guarantees about reliability.

Re:Singularity

[DamnStupidElf]

The only thing you're missing is support for arbitrary SIP-level proofs beyond type safety (e.g. support for arbitrary proofs of SIP behavior such as time/space complexity, halting, semantic properties, etc.) , and a formally verified self-verifying proof-checker to make sure the compiler is generating correct code and proofs. It looks like you're looking into PCC and TAL, so once you can ship the verifier with its own proof and self-verify during the boot process, you can be fairly certain that hardware errors are the only problem left. I assume you've already executing with a subset of the x86(_64) instruction set for easier verification. I figure that limiting code generation to the smallest set of opcodes can take advantage of formal verification done by Intel/AMD/others in processor design, while excluding all the complex protected-mode and virtualization instructions. Turning off SMM and injection of other arbitrary BIOS/EFI code would also be handy. The hardest part to model and prove correct will probably be the mutli-processor cache coherency behavior, but hopefully Intel at least has done some of that work already and can guarantee adherence to the specs.

Re:Singularity

[Anonymous Coward]

Type model has nothing to do with buffer overflows. NO such thing as a type safe assembler. Lack of memory sharing is NOT a good thing. Stop reading hype and learn to program.

UHmm ... and a non-hardware enforced protection is NOT protection of any kind, and you run it in the kernel address space? Thats the stupidest thing I ever heard. Nothing should run in the kernel address space except the kernel, and maybe not even that! (ie old Mach before they glued BSD into it).

Tricky

[JBMcB]

BSD style kernel
MacOS X multithreading
Solaris networking and filesystems
MacOS 9 system layout and management (auto install/remove via drag and drop)
Windows 7 start menu
System level support for IL - such as .NET or Java
Control strip from MacOS9
BeOS multimedia engine
Linux device drivers
AppleScript/REXX application scripting
OpenBSD code auditing standards, firewall
OpenVMS system partitioning, file versioning and backup

Re:Tricky

[Alomex]

Solaris networking and filesystems

Seriously dude? NFS is a security disaster.

Re:Tricky

[Anonymous Coward]

Maybe he meant the TCP/IP stack. I remember when Solaris 10 came out they talked about how the TCP/IP stack was redesigned to require as many locks which improved performance by 300%.

Re:Tricky

[Anonymous Coward]

Who said NFS? Maybe he was talking about ZFS.

Re:Tricky

[Anonymous Coward]

My personal list of wants. :)

Linux's kernel compile tool (keep only minimum modules required for my environment)
OS2 Warp's or even Netware's isolation of elements that have crashed while keeping the system running
HPFS's fragmentation handling
OSX's user level system config interface
OSX's user level application removal/installation interface
OS2 Warp's application and configuration files' grouping into segregated compartments
Windows 7's start menu for the 'all applications' section
Linux's device driver control over making them seperate or incorporated into the kernel
almost any POSIX's sys admin tools, file handling tools, and partition handling tools
VMS's built-in file version and backup
DataPoint's automated crash recovery ...

Re:Tricky

[JBMcB]

Definitely ZFS.

Re:Tricky

[Anonymous Coward]

+100 for REXX. Best OS command language ever! Too bad it never achieved wide adoption; I never knew a programmer who learned it and didn't love it.

You missed the task prioritization and network node aggregating abilities of QNX. Spread your application across a network and get simultaneous scaling and fault tolerance.

Android/PalmOS

[Anonymous Coward]

I would start with something simple, like PalmOS, only without any built-in corporation/government surveillance crap and without all the unnecessary bloatware. All an OS is supposed to do is load/manage applications. If I want to browse the web, I'll load a web browser. IT SHOULD NOT BE PART OF THE OS! There should be some very basic applications supplied (like a web browser and file manager) but the user should be able to supply their own if the built in ones are not to their liking. A nice small, efficient OS like this should boot up in only a few seconds.

The OS would be stored on a write-protected MicroSD making the OS itself (and basic functions like the default file-manager) immune to any kind of incidental malware infection. All drives would be heavily encrypted to help prevent data theft, either by criminals or criminal governments (the user would still be susceptible to the hammer unfortunately).

The device would have a physical write-protect switch on the outside of the device that could enable writing to the User drive for application installation and then switched back to protect mode for execution. The System drive would remain write-protected at all times. This should eliminate most virus attacks. If any applications (or any of its pieces) file size, and (CRC, or MD5, or checksum) change, it would be flagged as possibly infected. These three steps should reduce the amount of malware considerably. Of course there will always be social engineering (tricking someone into leaving the User drive write-enabled), or black bag jobs where physical access to the device is lost to the authorities for a period of time. It is just not possible to eliminate every threat. The major weakness to this would be content created by the user using an application. That would have to be stored/manipulated in a Data drive (separate from the User drive) which would always be write enabled.

All applications should install to one directory so if it is desired to delete that application, that one directory can be deleted and it all goes away. No more hunting for pieces here, there and everywhere.

Re:Android/PalmOS

[Beck_Neard]

As it so happens, we have an OS that's exactly like that. All it does it manage and load applications. Not only is the web browser completely outside the OS, the UI is largely outside of the OS as well. Of course, you need drivers and filesystems too, otherwise nothing works, but other than that it's pretty minimal. It's called Linux.

Re:Android/PalmOS

[damnbunni]

So, an Amiga with the OS on a floppy with the write-protect tab enabled, and the applications on an Iomega Bernoulli 150 meg drive with the write-protect tab sticking out?

Oh wait, encryption. Well, I'm sure there's something on Aminet.

Or MirageOS

[WittyName]

Cloud aware, built in scaling, the cluster is OS. Spin up new instances in a fraction of a second:

MirageOS is a library operating system that constructs unikernels for secure, high-performance network applications across a variety of cloud computing and mobile platforms. Code can be developed on a normal OS such as Linux or MacOS X, and then compiled into a fully-standalone, specialised unikernel that runs under the Xen hypervisor.

Since Xen powers most public cloud computing infrastructure such as Amazon EC2 or Rackspace, this lets your servers run more cheaply, securely and with finer control than with a full software stack.

MirageOS uses the OCaml language, with libraries that provide networking, storage and concurrency support that work under Unix during development, but become operating system drivers when being compiled for production deployment. The framework is fully event-driven, with no support for preemptive threading.

======= And ==========

What is Mirage OS?

Mirage is an exokernel (also called a Cloud Operating System) for constructing secure, high-performance network applications across a variety of cloud computing, embedded and mobile platforms. Mirage OS was initially designed to for cloud use, which is why we call it a Cloud Operating System. Mirage OS applications are developed in a high-level functional programming language (OCaml) on a desktop OS such as Linux or Mac OSX, and is then compiled into a fully-standalone, specialised microkernel. These microkernels run directly on Xen Project hypervisor APIs. Since the Xen Project powers most public clouds such as Amazon EC2, Rackspace Cloud, and many others, Mirage lets your servers run more cheaply, securely and faster in any Xen Project based cloud or hosting service.

Win95 UI + BSD/Linux OS on ZFS

[xdor]

Simple, non-nonsense interface, reliable ALT+TAB (damn you OS X!), good package manager, with all the redundant filesystem management wizardry. Good support for GPUs and the latest OpenGL/DirectX as well.

Re:Win95 UI + BSD/Linux OS on ZFS

[Anonymous Coward]

reliable ALT+TAB (damn you OS X!)

The OS X way (cmd+tab alternates applications, not each window) is the superior way.

Re:Win95 UI + BSD/Linux OS on ZFS

[xdor]

Yeah, I don't care what modifier it uses -- on OS X it doesn't freaking work!

Sometimes it goes to the previous app and sometimes it goes to the next app, and sometimes it goes it some other app purely at random.

I think Microsoft must own the IP for the "sequential movement between apps by key-stroke combination" and Steve Jobs was too irate to pay Bill for anything. So they just put a randomize into the routine to avoid a lawsuit.

Re:Win95 UI + BSD/Linux OS on ZFS

[Anonymous Coward]

YES! Sweet sweet validation. I've wondered if I was imagining that for years. There is no discernible logic to it.

Re:Win95 UI + BSD/Linux OS on ZFS

[Anonymous Coward]

Linux Mint 17.

Re:Win95 UI + BSD/Linux OS on ZFS

[Anonymous Coward]

Yeah, I don't care what modifier it uses -- on OS X it doesn't freaking work!

Sometimes it goes to the previous app and sometimes it goes to the next app, and sometimes it goes it some other app purely at random.

I don't know what you're smoking, but on every Mac I've ever used, command-tab goes to the next app, and command-shift-tab goes to the previous app.

The thing to remember is that switching to an app puts it first in the list, so if you hit command-tab and let go, then do it again, you're back where you started. Hold down the command key and keep hitting tab to get to the app you want. Dead simple.

Re:Win95 UI + BSD/Linux OS on ZFS

[Anonymous Coward]

Yep, win95 interface is about optimal. xp was pretty good, as was win7, but what came between and after... *shudder*.

Re:Win95 UI + BSD/Linux OS on ZFS

[thogard]

The sort order was changed in 10.6.0. It was refixed in 10.8.something but sometimes the order of the or windows apps changes and the then being able to hit two keys and know what is going to happen goes out the window. I've found it frustrtating enough to downgrade versions of OS X.

Re:Win95 UI + BSD/Linux OS on ZFS

[Anonymous Coward]

If you want a simple, non-nonsense UI, take Amiga Workbench. One thing that was always impressive with it was its responsiveness, something I seriosuly miss in machines that have a nearly thousandfold processing power.

If you moved your mouse, the pointer did move, regardless how busy the machine was, and menus and buttons were likewise quick to respond.

And yes, I know, there are a lot of things it could not do back then, but the speed which the Amiga UI exhibited is still unsurpassed.

And, if possible, add something like the AmigaOS "assiign" program. It allowed for virtual "drives", i.e. you could create an assignment "x:" pointing to a path or a set of paths, and you could then use the "x:" as part of a path in any application. There were global and local assigns, and the assignments designation was not limited to one letter like on windows (and it was not handles seperetely from the path as windows does it!). "assign" was such a powerful concept that it is difficult to explain all the (positively) interesting applications of it. It is something I really, really miss on Linux.

Re:Win95 UI + BSD/Linux OS on ZFS

[zennyboy]

I wish you'd logged in as that was a really well thought comment...

Re:Win95 UI + BSD/Linux OS on ZFS

[Anonymous Coward]

Well, they must have fixed it sooner than that, because I just booted my Macbook running 10.6.8 and Command-Tab works like it always did - the app I switch to becomes first in the list, and my previously used app follows right after it. All the rest stay in order and I never see a "random" app activated.

Re:Win95 UI + BSD/Linux OS on ZFS

[cbhacking]

Search on Win95 was nigh-worthless. Even back then, you had to waste a lot of time organizing stuff or you'd lose it utterly. A decent OS should (and some do) have search features that make this a non-issue.

Any time I try using a pre-Vista version of Windows - a blessedly rare event now, with XP out of extended support - it drives me insane. I can't launch programs from Start using search, I can't quickly find files across a folder tree using search, I have to spend a bunch of time navigating menus / directories even on a well-organized system, and visually scrolling anything else...

Win95's UI was minimalist, but it wasn't *good*. The abysmal search was only one of its problems. If you remember it fondly, I suspect you haven't used it in a long time.

DreamOS

[duke_cheetah2003]

Linux kernel. Microsoft's GUI and API compatibility. 'nuff said. Linux GUI is still way behind M$.

Re:DreamOS

[Anonymous Coward]

E19 is really far out. The sheer number of features and options is almost alienating but if you grasp the level of control and personalisation , you can't go back. Independent virtual desktop by screen is awesome.

Is the mouse is not your thing try i3 it's a totally different paradigm, a kind of tmux for x on steroids. But white E19 I can have a tiling WM on specific desktop..

vxWorks

[thinkwaitfast]

Real-time task management. That is all.

Re:vxWorks

[Anonymous Coward]

you must be a script driven slab of astroturf

it would be difficult to imagine, much less execute on, a more poorly
designed and byzantine set of system programming abstractions
than those in vxworks.

i worked there briefly and everyone was either suicidal or blind drunk
after being asked to add just one more crappy feature on that giant
shitpile of a thing day after day

you want reliable response in an embedded system, its considerably
easier to just write it yourself

praising vxworks, even minimally, is like calling North Korea a shining beacon of
government transparency and personal liberty

Easy

[Anonymous Coward]

Mac OS + Direct X

Done.

Lots more from AS400/OS400

[dwywit]

The hardware abstraction is a fantastic feature for a growing business - upgrade your hardware across different processor groups, and you don't have to re-compile your software.

OS400 has a compilable control language, and a command creator. Take your own utilities (equivalent to your favourite scripted/powershell jobs), compile them, then create a parameter-accepting command out of them, with optional menu-driven screens and context help.

DB/2. Not the best, but it's inbuilt, and accessible with system utilities/calls, using any language on the system, including control language.

Object-oriented - everything is an object, and you can't "accidentally" execute something that isn't a program.

Somewhat more sensible command abbreviations than *nix. For example, "display" commands always start with DSP, e.g. display system status is dspsyssts, display object is dspobj, etc.

You'd be correct if you said many operating systems have these features, but they've been present on the AS400 since 1988, and are very mature/stable. I live in hope that IBM or someone more talented than I will release OS400 for x86.

Re:Lots more from AS400/OS400

[serviscope_minor]

Object-oriented - everything is an object, and you can't "accidentally" execute something that isn't a program.

how do you intentionally execute something which isn't a program, e.g.:

ls blah | sed | blah | sh

Re:Lots more from AS400/OS400

[dwywit]

It's a bit hard to explain, and I really don't mean that in a patronising way. I guess I meant it in the way that you can't flag a non-executable file with permissions that could make it executable.

OS400/IBM i/whatever they call it now has a very robust object model and security system. A file can only be flagged "executable" when the system is happy to call it so, e.g. it's been successfully compiled. You can't just add ".exe" onto the end of a (for example) specially crafted mixed text+binary file in the hope of causing a problem, you can't patch binaries directly, and you have zero access to the binaries of the operating system, except of course via patches and upgrades sent from IBM.

My statement was a bit simplistic, sorry 'boot that.

Re:Lots more from AS400/OS400

[raftpeople]

Whether the attempt is intentional or accidental it's simply not allowed. Each object has X valid operations that it can perform, which eliminates all kinds of accidental and intentional (hacking type stuff) mismatches of objects and actions that can cause problems on the system. It's a very nice feature.

Re:Lots more from AS400/OS400

[raftpeople]

An additional item I would love to see in other systems that OS/400 does very well: Very detailed logging and error handling with detailed explanations and possible causes right in the text, much more detailed info than I have seen on any other system, a made up example which doesn't really show how good it is but gives a flavor for the style of every single error message that could come up:
Msg 123456 - Unable to open file ABC
Reason Code - 17
Possible reasons:
01 - File does not exist, 02 - User does not have permission, ... 17 - File is exclusively locked and cannot be shared, ...

Re:Lots more from AS400/OS400

[400_guru]

Because of the object oriented nature IBM i (the modern version of OS/400) a thing is either a program or its not. Creating programs from source generates an intermediate level code. This code is then compiled into lower level code that differs by O/S version and by the underlying hardware (e.g. POWER8 or POWER7). Because the object is a program you may only perform actions allowed on programs. Execute it, rename it, save it, restore it, move it, delete it, change it's owner or authority. That's it. You cannot change it's type to say 'File' and then edit it and then change it back to 'program'. The O/S blocks that which is beautiful and it's why there are no anti-virus products for IBM i! You also cannot execute something that is not a program. So that avenue of attack is also unavailable.

Re:Lots more from AS400/OS400

[400_guru]

DB/2. Not the best, but it's inbuilt, and accessible with system utilities/calls, using any language on the system, including control language.

Say what you will but DB/2 on IBM i is the most SQL compliant database of any. It scales well and being built into the system means tuning it is not often needed. In addition there is near zero database administrative effort required for most IBM i systems. Since all disk space is typically viewed as a single store, space for tables and collections is automatically drawn from and returned to that space.

For desktop OS, I'd tale BeOS' responsive handling

[blind biker]

BeOS has an incredibly responsive UI. I am not a software engineer so I am not sure which part of the OS is in charge of this, but it's something no other OS has been able to do, before and since: be perfectly responsive to user commands (keyboard and mouseclicks). What this means is: no matter what the computer is doing at any given time, the UI will react to the user commands. There is no file-copy too big, a computational task too complex, that the reaction to a user's command would be delayed. BeOS has spoiled me so much, because with that OS, user comes first, always.

Re:For desktop OS, I'd tale BeOS' responsive handl

[Anonymous Coward]

While it is true that the BeOS is an incredibly responsive OS, it should be noted that Atari's TOS (ST/TT/Falcon) era also had that kind of immediate responsiveness. It's incredibly important, and something that nobody has gotten correct outside of Atari and Be.

Unix Haters Handbook ?

[Anonymous Coward]

I'm really surprised no one brought this up

Re:Unix Haters Handbook ?

[weav]

TOPS-20 FTW! Now about 36-bit-word addressing on the x86....

Re:Unix Haters Handbook ?

[Evan Langlois]

It was originally 20 bit

Totally this... what else could you possibly want?

[BevanFindlay]

• The multitasking prowess of DOS.
• Ribbon interfaces on everything.
• The beautiful colour palette and icons of Windows 3.1.
• The stability and driver handling of Windows ME.
• The simplicity of configuring drivers of Windows NT.
• The memory footprint of Windows Vista.

The sad thing is that I was trying to think of a variety of examples, and they were all from Microsoft. Hmm.

Is that a new distro

[Anonymous Coward]

FrankenOS? Which distro is that?

FreeBSDindows

[Anonymous Coward]

FreeBSD everything except Windows APIs so I can actually game.

Re:For desktop OS, I'd tale BeOS' responsive handl

[Anonymous Coward]

In addition, the BeOS filesystem rocked too. Opening any folder was instant. Searching the entire drive was instant, and live! Changes to the filesystem would immediately be reflected in the search window.

VS OS

[jmccue]

Wang VS OS with X, but that OS was so unique it really would have been impossible to port X to it, or almost anything from a UN*X system for that matter. So I would just take it ported to x86_64

Religious Questions

[gavron]

TL;DR summary - at the end of the day the gamer wants a tight Windows system. The server admin wants a tight server (LAMP, WAMP, XAMP, MEAN, etc.). The hardware developer wants all the latest drivers. The R/T guys want predictable and repeatable, and Donald Trump wants it not to be produced in Mexico. However, that's not what OP asked about, and his interesting article relates to features on clustered filesystems that are cool to have and not available outside of the [really out of date/obsolete] OpenVMS.

Long Version:
Whenever someone asks about "best" OS or app or features to have in one... invariably it leads to the proponents of Windows, Linux, FreeBSD, MacOS, (and kudos to the poster who brought up BeOS!) etc. all jumping to extoll their virtues.

Ironically the OP asked about OS and everybody jumped into talking about monolithic kernels... filesystems, and only a couple discussing other elements of the OS which is queue system (OpenVMS really had that one sewn so tight it was awesome).

Interestingly tho the original ARTICLE talked about a clustered filesystem environment. It would appear OP is right on this one - only VMS did it. Some of the functionality for single-host stuff is now beat by BTRFS, but the clustered writeback, locking, and other features mentioned in the PDF are without compare in anything else today

Ehud
Tucson AZ US

Re:Religious Questions

[thinkwaitfast]

OT, off the wall question... were you at LAMPF in '89?

Re:Religious Questions

[gavron]

I was - a LAMPF DAC system manager.

Not to be confused with my dad (same last name) who is a physicist :)

Ehud
P.S. And you are? :)

Re:Religious Questions

[thinkwaitfast]

As was I (physicist)(in training). Tequila sponge. I only vaguely remember that period in my life so don't even recall if I knew you other than through the computer.

Re:Religious Questions

[gavron]

Well if I knew your name I would try and recall myself. It's part of my "forgotten years" as well... all went by so quickly ... and so long ago :)

Ehud Gavron
first dot last at gmail

Re:For desktop OS, I'd tale BeOS' responsive handl

[mc6809e]

Sorry, but BeOS is in an entirely different class than is TOS.

Atari's TOS was quick and responsive because it was simple, lacking many features we take for granted today like preemptive multitasking and multi processor support.

BeOS was responsive because it was a complex, full featured OS that was also well thought out and well designed.

Re:For desktop OS, I'd tale BeOS' responsive handl

[Anonymous Coward]

Good OS, RIP.

drivers!

[gl4ss]

don't forget being able to use drivers made for win X and osX...

Boot up time

[jpatters]

I just want it go go from off (not standby) to an open document in a useful application in under three seconds.

Re:Boot up time

[Anonymous Coward]

This is limited by size of OS + "useful application" and disk transfer access time and transfer throughput and then CPU performance.

Most hard disk systems could do this if the combined size were about 50-100 MB.
SSDs might bump this up to the 300-800MB range.

And that's why only a stripped OS like ChromeOS can boot that fast... it's the simplest OS we have today.

Re:Boot up time

[jones_supa]

Windows 8 and newer come close to that, as they start up part of the OS from an hibernation image.

Re: Sorry but your post's off topic... apk

[Anonymous Coward]

Binary seek requires sorted the data. So what happens when you need to insert an element? You must sort the entire set again. You can mitigate this by leaving giant gaps, wasting memory.

If you're gonna sort it all anyway on update, you may as well use a perfect hash. There are perfect hash function generators which can compute a minimal perfect hash function for millions of keys in just a matter of a few seconds. Then you get O(1) access, no collisions, and no holes.

Of course this means updates have much higher latency. Jitter is one reason why people use trees and linked lists. They're not the fastest, but they're never slow and always efficient algorithmically.

Too Young

[Anonymous Coward]

But I'd like to see a LispOS in action.

For some reason, I think computers systems should be a lot smarter and more flexible than they actually are. Instead, I often find a computer program that does 90% of what I want, with no way to get the other 10% done other than by rote tasking by myself.... or in case of guis, they are pretty much only good for what they were doing 30 years ago, just a bit more refined.... but they take no tasks easily... like "rename these 30 files x*** or y***". The only real upgrade is they can handle bells and whistles like mp3 files or video, etc.

parts

[JohnVanVliet]

NONE!!!!
Frankenstein'ed installs are to BE AVOIDED!!!! at all costs!!!!!!

if you want a custom install
Linux From Scratch
or
Slackware
or
Arch

there is no way in hell you can keep a Frankenstein'ed install up to date with updates

IEBGENR instead of cp, of course

[Anonymous Coward]

JCL really allows you to get down and dirty with the metal and specify exactly what you want. What cylinders to store that file on, whether there's blocking. Even useful stuff like whether the operator should put a ring in. (no kiddies, that's not putting a ring on it a'la Beyonce)

Re:IEBGENR instead of cp, of course

[davester666]

It's OK if you don't want to go steady with, or [oh, the horror] marry, your data.

Don't bring up VMS please

[Anonymous Coward]

I work for one of the only companies that still use VMS literally everyday for production. Don't you ever give it credit for anything it is an unholy mess of shit ware. The clustered file system you speak of while it will solve your problems of file access is ungodly slow and is actually what our processes sit doing 99% of the time. We have calculations that take hours because of IO when the actual calculations only take mere seconds or minutes

Re:Don't bring up VMS please

[0xdeaddead]

Are you still on VAXen? I thought the Itanium stuff was at least newer than the Dec Alpha...

My choice?

[thinkingrodent]

Sad as it is, I'm happy with a nice, vanilla XFCE 4.12 Linux distro with a 4.x kernel.

Recognizing Pictures

[Anonymous Coward]

Where is the file system that can just scan vids and pics and assign metadata like "girl in blue dress" or "white truck" and let me find any file by using human expressions without requiring the user to fill in the blanks?

Google Photos does that now. It has to rely on the machine learning system running in the cloud though, so your device has to connect to the internet.

rip oskit

[0xdeaddead]

http://www.cs.utah.edu/flux/os...

so cool back in the day, but it didn't really take hold for some reason. too bad.

Virtualize it all

[Anonymous Coward]

Just make it easy to use a hundred VMs, and to move things like files between them. Make sure the system can auto-detect what OSes are supported when you try to install software, so that if I d/l a copy of GIMP for OSX from inside a Windows 7 VM, it will load the correct VM.

Microsoft could easily do this, but their marketing department is obsessed with moving everyone to the newest version of Windows.

I'm not a creative guy

[OrangeTide]

I'd use all of the parts of Plan 9, then stop there.

It would have an Amiga Mother board

[Trax3001BBS]

True multitasking, frankenos as it's pry not possible to gather all of the patents and OS to ever run it again.

Re:It would have an Amiga Mother board

[fuzzywig]

Also, the Amiga had the best case handling in a file system:
Say you have a file called File1 (no need for an extension). You could access the file with any capitalisation you like, file1, FILE1 File1, FiLe1 or whatever takes your fancy.

However, if you were mad, you could also have two files in the same directory names FILE1 and file1 (or any other set of characters).

It had/has the benefits of a case sensitive filesystem, with the ease of use of a case-insensitive one.
Now, if someone could just explain to me why case sensitivity is important in a file system...

Re:It would have an Amiga Mother board

[Trax3001BBS]

Also, the Amiga had the best case handling in a file system:

Say you have a file called File1 (no need for an extension). You could access the file with any capitalisation you like, file1, FILE1 File1, FiLe1 or whatever takes your fancy.

It had/has the benefits of a case sensitive filesystem, with the ease of use of a case-insensitive one.

Now, if someone could just explain to me why case sensitivity is important in a file system...

And why to this day I write the file name "HOSTS" in caps as it didn't understand it unless it was, until version 3.

Genode and Nix

[Anonymous Coward]

I really wanted something like Genode's micro-kernel and robust security and process isolation along with a pure functional package manager like Nix. Then Genode basically did that. I'm so excited. The other thing I wan't is a way to use it as my host OS and virtualize my existing windows stuff on it. That apparently working now too. I just need real VGA passthrough support (for decent windows gaming perf), then I'm buying hardware with IOMMU support and setting up my ideal frakenos. It's so close!

A real grab bag

[Anonymous Coward]

OS/400 for the database, Linux for the communication layer, Windows for the graphical interface.

formal verification

[Anonymous Coward]

Package manager: Nix
Kernel: seL4
Default compiler: compcert
Userland written in Haskell

Genode, Plan 9, and OpenBSD also have some interesting bits worth scavenging

Tandem Guardian.

[jcr]

Reliable systems have existed in the past. Windows is so bad, that people think of UNIX as reliable.

-jcr

Re:Tandem Guardian.

[stardaemon]

Still exist, even, if somewhat expensive. Though it's HP NonStop rather than Tandem nowadays.
https://en.wikipedia.org/wiki/...

I work with that system on a daily basis.

Hmm

[behrooz0az]

I'd use Firefox as OS and chromium as browser.

MS-DOS

[maroberts]

for its memory management.

I'll get my coat...

Parts of AmigaOS of course

[johannesg]

Assigns (randomly named identifiers for things in the file system). ARexx (fully integrated, comprehensive scripting support, don't care about the language, throughout the OS and all applications). Screens (the ability to group windows and more importantly tasks together). A design that doesn't require constant disk access and thus remains responsive at all times. I'd also choose any windowing environment that is not X11, and since we're talking Amiga anyway I'll choose Intuition (a sane windowing environment).

You can sort-of pretend all of these things exist in Windows or GNU/systemd, but in reality they are pale imitations of the original. Screens for example - that works because applications have knowledge of them, and use them intelligently, not because you just happen to be able to assign windows to a workspace manually.

Re:Parts of AmigaOS of course

[Anonymous Coward]

I agree with this, to add some of my own thoughts it'd also would have AmigaOS locale-system, it's still the best I have encountered for having programs in different languages. The Workbench would have to be updated quite a bit, the icon-system with it's .info-files feels a little archaic even if I sort of like it, but often the contents of a drawer (directory for all you non amiga-users) was chaotic and you had to spend way too much time just snapshotting windows to have everything look nice.

From other OS:es I would like a:
* Shell with bash (or similiar) with all the common GNU-utilities
* Memory protection
* FUSE
* Some sort of multiuser support

And probably lots of other things I have forgotten...

Re:Parts of AmigaOS of course

[hogne]

You beat me to it. We've come a long way already: http://friendos.com/ - it has an AREXX equivalent (Dormant), Screens and even object oriented C (like BOOPSI) for those who feel adventurous. It's based on web technologies, so by using webkit on the below layer, you get to use it as an X session - and if that's not your cup of tea, it has its own server application, so you can distribute it online as a web service. It allows you to program services, modules and apps on the server side in any language (native or scripting), and gives you a full GUI framework for making sandboxed, secure web apps (that becomes responsive, so you get it on your TV and phone as well). Oh, and the server application connects to other server cores, so you get a distributed "macro computer" :)

the best of it all

[Anonymous Coward]

two words.. trumpet winsock!

it's already here!

[crhylove]

I can't think of many improvements that would be possible to Linux mint Debian edition. Though it would be cool if all my devices shared all resources real time I guess. Oh, and I'd like my various desktops to have different wall papers.

Re:it's already here!

[Evan Langlois]

Don't know what you mean by sharing resources, but you CAN have different wallpapers for each desktop. Compiz will do it. And you could probably write a Gnome shell extension for it (they are written in Javascript).

Re:it's already here!

[crhylove]

I meant sharing bandwidth, CPU cycles, memory.... Like I'd like all my devices to be one big super computer with local interfaces. So I could play GTA on my weak android tablet, for instance. I'm always frustrated at the slowness of my old work laptop when it's less than 2 feet away from my brand new i7 with 16gb of ram.

Re:it's already here!

[crhylove]

Oh, and I don't do compiz because that breaks xrdp. I bet there's a way to get separate desktops in Mate, but I didn't see an easy option and don't really have time to track down and hack largely cosmetic issues. :/

Re:it's already here!

[Evan Langlois]

You could try clustered computing. But that requires specific application support in order to be efficient. There are certain overheads. As for playing games on your tablet, you could probably hack that up. Steam is working on some technology to do similar things. You basically just run the game on one device and then compress the pixels and shove them out the network card - not much different than streaming a movie. As for your old work laptop ... just use the new one! You could repurpose it by loading Linux and a desktop designed for older machines such as XFCE. Basically, you aren't going to be able to just share everything. There are barriers to scalability and connectivity.

Easy

[nyet]

Systemd, avahi-daemon, pulseaudio, networkmanager. What an awesome combination of great software!

File System

[camperdave]

There were two things I liked about the VAX/VMS filesystem that I would port over. The first is version numbering. A file would start as filename.1 and the number would increment with each edit/save cycle.

The second thing would be logicals and overlayed directories. They worked like a stack of transparencies like the human anatomy entry in an encyclopedia. The base layer would be a read only version of the operating system. Above that would be a writable layer. Above that, for development users, would be test versions of new OS elements. Regular users wouldn't see these layers. Above that would be applications (read only) with a writeable layer above it.

The purpose of the writeable layer over the read only layer is to trap attempts to overwrite system files.

Re:File System

[Anonymous Coward]

When you start making lists of things like "The base layer would be a read only version of the operating system. Above that would be a writable layer. Above that, for development users", it's a sure sign that you have the wrong design, and what you should have is a configurable list or set of things, in this case, a configurable list of stackable filesystem overlays.

The problem with most OS, including unix, is all the cruft that people decided to make a special case for performance, efficiency, "because this thing is special" etc, that gets in the way of getting the job dones and proliferates the number of tools required to configure, operate and even build the system.

My idea of the perfect FrankenOS, is one which takes components from other OS and gives them uniform interfaces ala Smalltalk, Lisp or Python.

Hierarchical Filesystems in unix are actually a bad abstraction, here's why: They define a directed (semi)-acyclicgraph of nodes which contain a capricious set of fixed meta-data (less fixed with ACLs and Posix extended attributes) and then inside those nodes are a bunch more data structures, but the punch in the nuts, is that the data structures are linearised in an array, so every program/programmer has to adopt yet another system/library to access the internals of those objects. And you lose all the functionality the ACLs, permission bits, uid/gid, and extended attributes gave you over the files for all the nodes beneath this arbitrary node level called a "file".

My ideal system would combine the efficiency of files (domain specific encoding) with the generality of object graphs, by allowing prototypes to be declared for arbitrary levels of the graph that define the transform from the ondisk/onflash/etc format to the programmers object representation in the form of a grammar.

Also my ideal system would have per process namespaces like Plan9, but I would discard the 9P aspect as too imperformant.

With these two features it would subsume everything you described as mere operating specifics. Generality kicks ass.

Re:File System

[Pascal Sartoretti]

The second thing would be logicals and overlayed directories. They worked like a stack of transparencies like the human anatomy entry in an encyclopedia. The base layer would be a read only version of the operating system. Above that would be a writable layer. Above that, for development users, would be test versions of new OS elements. Regular users wouldn't see these layers. Above that would be applications (read only) with a writeable layer above it. The purpose of the writeable layer over the read only layer is to trap attempts to overwrite system files.

Isn't it what UnionFS offers ?

Re:Try and make an OS that viruses couldn't target

[Drinking Bleach]

I'd add the ability to run Windows binaries in emulators, but they can't access other programs than themselves. If that was a problem, add a phantom disk image so it could see other files that you place in the phantom disk image. Imagine each Windows emulated program saw their own personal c:/ , and it and you can populate it with files.

So... Wine with a new WINEPREFIX for each program?

I figure if the software you download can't get out of the Windows emulator or its own personal filesystem, it can't mess with your OS or the rest of your filesystem. If it can't record your keystrokes unless you have the window actively open, a keylogger can't get you either. The problem is that we probably don't have perfect Windows emulation. Another problem is you have to be able to trust your drivers or that is a possible vector to an attack.

Run Wine in a Docker image? That's pretty well-sandboxed. and easy to set up.

Frankensteinian, eh?

[Anonymous Coward]

Windows stability with security patches issued by Adobe flash developers, C64 loading times, iOS openess, Choice of Unity or Metro GUI, Cisco cli, Linux game support and BSD vitality.

The only hardware supported would of course be inspired by a crossover of Nokia N-Gage and Nintendo Virtual boy and use mandated by national policy.

Re:Totally this... what else could you possibly wa

[freeze128]

And yet you *didn't* choose to integrate a web browser into your OS. Hmm....

Re:Try and make an OS that viruses couldn't target

[dcollins117]

Can you explain how a keylogger in Windows gets access to user input in another application without a security approval by the user when it was installed?

If I remember correctly there's a system-wide hook chain for keyboard events, and you add the logger to the chain with a system call. SetWindowsHookEx or something like that. I'm not really sure what the purpose of that function is except to install key loggers, but perhaps someone has a legitimate need for it.

As far as getting an application to run after rebooting, there are (unfortunately) a myriad of ways to go about that too.

Caveat: it's been a number of years since I used Windows so this may have changed in more recent versions.

Mine

[Anonymous Coward]

What I'd take:

From Windows:
- The Taskbar
- The Start Page (without the Metro Apps)
- Window snap-to/drag-snap
- The Full Backing/Support of hardware vendors/product vendors/games/dev environments
- CCleaner
- Driver Support
- The Windows XP Background

From various Linux OSs
- Bash
- Multiple Desktops (see also: Mac)
- apt-get
- Free/open source software

From Mac
- The multi-touch touchpad
- Multiple Desktops (see also: Linux)
- Photoshop/UX product support
- Scroll on mouseover

What I'd leave:
- Notepad (Windows)
- Internet Explorer (Windows)
- Metro Apps (Windows)
- The (sometimes annoyingly toxic) open source/freeware community (Linux)
- The flaky crash-prone interoperability of certain programs (Linux)
- The 'elite-ness' or 'fanboyism' (Mac)
- The price tag (Mac)

Something Hybrid

[os2fan]

GUI

The OS/2 and Windows shells together, complement features missing in the other. The sort of hacks you can do with REXX and the WPS don't cut it in Windows, but MSFT got the idea right when they had proper files, rather than just EAs for their shell. I even added items to a windows desktop remotely. IBM gave a lot of flexibility to the REXX api to the shell, but the SETUP string was a single element, and it could had been something akin to an environment in an INI file.

One could had done some interesting things with PIF files. For example, they could had launched an application off-path, and the same icon could have had several pages for different operating systems, so for example, the Boxer.pif could launch boxer, tko or boxer/2 for different operating systems. You put a mob of these in a directory and you could launch different applications without having to go to the desktop, or some menuing system.

So, you could have a fairly decent file patcher, and still just have a single link in the PIF file.

EXTPROC in OS/2, is the DOS version of the UNIX /! thing. Where in UNIX, you have to have perl in /usr/bin/perl or something, in the new shell, it would simply look for perl.pif, either in the pif or extproc directory. Doing this means you could run the same perl script under different environments. Extproc could be added to other languages, so one could see or launch a script under something like FAR or some other program that takes an external script.

For example, the OS/2 EWS "StartDOS" expects a REXX script as input, so if you made REXX handle EXTPROC, you could start the startDOS script as extproc /rexx StartDOS

Alternate DOS/Unix command line personalities, so that programs that look for UNIX or DOS would not have to be rewritten. Likewise, you could convert the drives into different pointers in the unix system, so that you could have multiple cwd's (eg subst c: /usr )

Run a file under a different extension. You can already do things like start OEMSETUP.BIN with the extension, and many utilities load .TMP, .MOD etc as exe files, so it's hardly a bug there. Note that in Windows 98+, the xcopy32.mod is just xcopy.exe, and works without the other files if so loaded. What it can do is to allow you to open an exe as a rar file, etc, eg open .rar rarexe.exe would instruct RAR to open it as a rar file.

Clipboard and Select interface from the command line. You could do things like 'select = ls pk*.*' , which would add the output files of the ls command to the select list, select /a is append to the list, and then /copy. /paste. /cut, /all, /move /reverse, /u would do various things to the list. @select is then a virtual list that can be used in any command that accepts lists.

Multiple desktops at session level. Something like LaTeX involves lots of little files in the path, which are largely used purely for Latex. A TeX desktop (by setting desktop=tex in the PIF), might open up a window of utilities that share more or less like the same virtual machine. You create objects in there which inherit the common Latex settings, so ye could have an Editor in there, like WinEdit, etc. When you are not playing LaTeX, these are no longer in the path. These could be in 'groups' like Win31 Progman, and you put links out to the common paths.

Multi-language interface, so one can use REXX. Lua, or a variety of other languages to access system functions. Something like 'fdisk' might report things like the size and file system of a device, and a function-call like call 'fdisk' diskfree, 'c:' could be used in any script, including command.com.

Network as a separate desktop, so one can be logged into several networks at once.

Re:He's a decent enough young man... apk

[Anonymous Coward]

Translation: I am mentally ill and need help, but the nature of my illness prevents me from realising this. I'm so ill, in fact, that I think it's a good idea to advertise my anti-advert software in such a fashion which can't be defeated by my anti-advert software, and fail to see the irony. I will stalk and harass those who point out problems in my approach to software, advertising and mental health, to the point of alienating even more and more people, ensuring the cycle of my mental destruction will continue unabated.

I'd change not the OS, but all my games

[Anonymous Coward]

Linux is fine, but I still use Windows because so many games require it and run better on it. But I miss having configs in text files in git. So trivial to set things up.

It seems however, that this is now less of a problem than it used to be, and with SteamOS it'll probably be even less significant in future.

Odds and ends

[NorthWay]

A versioning filesystem (VMS still seems to be the only one around).
Datatypes Next-Gen (Amiga style, but handling streaming etc).
SASOS.
Proper multiscreen (in the Amiga sense) support. Happily with dragging and overlapping.
A smart shell that knows command options, behaviour and their documentation, like I believe AS/400 does.
Devices like the Amiga has - "FTP:" is a filesystem, there are editors that export their current files through a filesystem device, and many many more.
A framework for handling compressed data like XPK for the Amiga.

Re:Odds and ends

[hogne]

Again, this is what we're developing. At http://friendos.com/ you can take a look for yourself. We've taken the best features of AmigaOS and blended them in with the best from Linux desktop environments. When it's completed, you'll have a desktop environment that is responsive, working on all your devices (even your smart TV) - and with an option of running it instead of Gnome. :)

Re:For desktop OS, I'd tale BeOS' responsive handl

[Anonymous Coward]

BeOS is essentially a super-tweaked AmigaOS...
It is also incredibly efficient...

I kinda miss AmigaOS 3.1 w/ Picasso GFX handling, and a decent amount of memory. By decent, I mean a 1GB or so...
Someone else name a proper multi-tasking OS that will work with 1 MB from a floppy that is actually usable.;

Obvious design feature

[dbIII]

(linux's swap partition is bonkers).

It's so you can put it on a different hard disk than the one you want quick access to.
And a swap file is very easy and has even been an option on install with most distros for close to a decade. Fedora Core 1 had that option FFS, now we are up to Fedora 20+.

DRS/NX

[Spankalot]

Had it all in the 80's. Multitasking, networked.. Just made by a company that couldn't market water in a desert.

Re:For desktop OS, I'd tale BeOS' responsive handl

[DNS-and-BIND]

BeOS' problems weren't technical, that's for sure. They were purely created by evil humans. BeOS should have won, but didn't, and for a very good reason.

"I once preached peaceful coexistence with Windows. You may laugh at my expense - I deserve it."
-- Jean-Louis Gassee, CEO Be, Inc.

System VI

[Roger W Moore]

System V was introduced in 1983.

The reason there has not been a new version released since then is because System VI couldn't run emacs.

Keep the VMS Help system

[Daemonic]

Type "Help", and get help! Comprehensive, hierarchical, actual help. Awesome.

Re:Keep the VMS Help system

[swb]

Is that an operating system feature, or just a historical case of a system modern enough to have online help but old school enough to have decent, comprehensive documentation of functionality and error messages?

The long term trend definitely seems to be that documentation is an afterthought at best, with a lot of things, especially errors, being totally undocumented. You might get lucky and find a decent O'Reilly-type book or roll the dice with some 900 page Sybex monster that's 80% screenshots of obvious GUI tasks.

Re:Keep the VMS Help system

[Daemonic]

I gather (from other comments in this thread) that the O/S did some of the command/parameter parsing for you, so I'd be inclined to argue that it was part of the O/S.

Now that I think about it, I think you could register your programs with the help system, and help for your program would automatically get included if someone just typed "help".

AmigaOS mixed with Linux and NEXT

[hogne]

AmigaOS had some amazing technologies - like AREXX and its intuitive file structure. It also had a very easy to use DOS implementation. NEXT was early on into network filesystems - in an integrated way. Linux is amazing for developers - it gives you access to everything right away. Put together the versatility and developer friendliness of Linux - the intuitiveness and user friendly AmigaOS, and the innovative NEXT, and you've got it (at least for me). In fact, I've already scraped together a team to build this new OS - and our team has come about 60% of the way to meet the December release date. To become easy to distribute, we based the OS on web technologies. This way you can use it both in the browser, and as a Linux / BSD X session. Look here - we just went public: http://friendos.com/

Beos/Haiku+AmigaDOS+OpenSolaris+VMS+Osx+Ubuntu

[An dochasac]

Start with the OpenSolaris rock stable (since the early 1992-2010?) ABI, add ZFS for its efficient support of flash, snapshots, encryption, RAID-z... Add BEOS/Haiku's user-meta-data indexed filesystem and AmigaDOS's backward/forward linked file allocation "table" (To turn off the computer, you switch it off, no fscking "start->shutdown" nonsense.) Graft OSX's time-machine onto ZFS's efficient copy-on-write snapshots for an improvement on VMS's auto-versioning files. Use Ubuntu's package manager, huge application repository and ATK accessibility features. Glitz it up with OSX's Quartz extreme GUI.

Re:Beos/Haiku+AmigaDOS+OpenSolaris+VMS+Osx+Ubuntu

[hogne]

Agreed. We've got so many good "experiments" over the decades that it's time to try a new OS formula. Linux is stuck in its crazy and cluttered, non standardized unix file structure. No scheme will fix this, unless it abstracts the file structure completely and gives the user an alternative. Gobolinux does a good job. But I don't know if they could go all the way. Mac tries, as it has inherited the unix structure too, but only for the first mile. IMO, Mac OS X was kind of abandoned for iOS. So it's basically the same as in 2001 - at least superficially. Windows is going crazy. Do they know where they are going? I think a completely new OS is the way to go. But based on our existing systems, for the drivers. Then put a layer on top that accesses all the resources on the web. :)

There IS no "perfect hash"... apk

[Anonymous Coward]

See subject: & no "1 size fits all" hash function for everything for starters.

You WILL also encounter memory fragmentations eventually like you do with linked lists using hashing. Not even unbalanced trees usage prevents this. Pointer hell and memory location all over RAM guarantee this. Exchange Servers PROVE it.

What if I don't required insertion? I don't in hosts (which is strong in my examples here) and download already sorted lists DEFEATING your "single objection" right off the bat.

* It all depends on application.

The fact here is that binary methods are MORE PERFECT in that they ALWAYS WORK in a std. fashion, for any list & guarantee steady solid performance as well as higher all-around "perfect hash"-ness, better than "perfect hashes" (no such thing exists) do for more things.

You can downmod this ALL DAY as you did before here -> http://tech.slashdot.org/comme...

Doesn't change a thing. I'm right as rain...

APK

P.S.=> Hashes are NEVER slow? Untrue. They're SLOWER TO DEVELOP (& again, aren't as universal as binary methods) in the first place, & are not as standard as binary methods are across MANY applications (again, why I try to stick by them - it's less overall work & std.)... apk

Almost no parts

[Anonymous Coward]

I'd use almost none of the existing parts except for some good file system. The key problem of all OSs I know is the handling of data persistency. They don't distinguish between volatile information (across time, versions and operating systems) and information that represents user choices and user-generated content. That's why its such a fucking pain in the ass to migrate systems. Even with smart migration assistants you end up readjusting everything and manually deleting or copying settings when switching to a new machine. The same for backups. You face a choice between keeping everything and then manually choose which of the 'secret hidden file sauce' will work and be needed in the new environment, or you only keep your own documents, loosing thousands of application settings.

There is no technical reason for this limitation, it just that existing operating systems do not reliably track meta data for the purpose of given data. Implementing this requires changes in all application (via some mandatory preference/data tagging API), so practically no existing applications can be used in the new OS except very low level stuff.

FrankenOS V0.1

[gedeco]

Since Frankenstein does have a more negative sound, my FrankenOS should definitely contain

- Cooperative Multitasking
- BSOD screens
- VI Editor as the only one allowed (Due to the complexity of the interface)
- Novell ipx/spx Netware for communications
- Floppy disk in case you need to perform manual updates

Hmmm looks like a sound description of WFW with Novell Netware connection.

What I post's nonsense dave420?

[Anonymous Coward]

"I just reply to you when I see you spamming Slashdot with your nonsense"- by dave420 (699308) on Friday June 19, 2015 @10:31AM (#49945047)

Why'd you agree w/ my points on hosts then? Quoting you:

"I'm not denying all those things" - by dave420 (699308) on Wednesday September 17, 2014 @11:39AM (#47927435) FROM -> http://yro.slashdot.org/commen...

Of course not: It's impossible to dispute HOSTS FILES superiority to other methods!

Since my points in favor of hosts SINGLE FILE native kernelmode faster part show hosts doing more w/ less vs. so-called 'competitors' many part messagepassing + cpu/ram use overheads laden slower usermode FAR MORE COMPLEX 'solutions' doing less than hosts do for more security, speed, reliability, + anonymity!

---

* QUESTION:

DO YOU WORK FOR AN ADVERTISING FIRM, or ARE YOU A WEBMASTER/WEBCODER http://slashdot.org/comments.p... , or a MALWARE MAKER, or ARE YOU AFFILIATED WITH 1 OF MY COMPETITORS?

Answer it!

As per your usual you'll avoid every question, or lie & You've been EXPOSED in your "motives" in the last link just above, lol!

---

"I'm simply pointing out that it takes an AdBlocker to block your spamming"- by dave420 (699308) on Friday June 19, 2015 @10:31AM (#49945047)

I bother you? Then WHY DON'T YOU DO IT & use 'em? Answer that!

(You stalk/harass me instead!)

OBVIOUSLY you don't & you're a "ne'er-do-well" troll & you have "other motivations" (the QUESTION above you'll AVOID TO NO END, "Gosh, I wonder WHY?" (not!)):

---

I make creating a superior more efficient solution EASIER (That's more than a mere trolling stalking harassing "ne'er-do-well" like yourself could *EVER* manage).

APK

P.S.=> See Dave420 the "pot puffing clown" SQUIRM - evasions galore will ensue (as well as effete downmods via sockpuppets to *try* vainly "hide it" -> http://slashdot.org/comments.p... )... apk

What I post's nonsense dave420?

[Anonymous Coward]

"I just reply to you when I see you spamming Slashdot with your nonsense"- by dave420 (699308) on Friday June 19, 2015 @10:31AM (#49945047)

Why'd you agree w/ my points on hosts then? Quoting you:

"I'm not denying all those things" - by dave420 (699308) on Wednesday September 17, 2014 @11:39AM (#47927435) FROM -> http://yro.slashdot.org/commen...

Of course not: It's impossible to dispute HOSTS FILES superiority to other methods!

Since my points in favor of hosts SINGLE FILE native kernelmode faster part show hosts doing more w/ less vs. so-called 'competitors' many part messagepassing + cpu/ram use overheads laden slower usermode FAR MORE COMPLEX 'solutions' doing less than hosts do for more security, speed, reliability, + anonymity!

---

* QUESTION:

DO YOU WORK FOR AN ADVERTISING FIRM, or ARE YOU A WEBMASTER/WEBCODER http://slashdot.org/comments.p... , or a MALWARE MAKER, or ARE YOU AFFILIATED WITH 1 OF MY COMPETITORS?

Answer it!

As per your usual you'll avoid every question, or lie & You've been EXPOSED in your "motives" in the last link just above, lol!

---

"I'm simply pointing out that it takes an AdBlocker to block your spamming"- by dave420 (699308) on Friday June 19, 2015 @10:31AM (#49945047)

I bother you? Then WHY DON'T YOU DO IT & use 'em? Answer that!

(You stalk/harass me instead!)

OBVIOUSLY you don't & you're a "ne'er-do-well" troll & you have "other motivations" (the QUESTION above you'll AVOID TO NO END, "Gosh, I wonder WHY?" (not!)):

---

I make creating a superior more efficient solution EASIER (That's more than a mere trolling stalking harassing "ne'er-do-well" like yourself could *EVER* manage).

APK

P.S.=> See Dave420 the "pot puffing clown" SQUIRM - evasions galore will ensue (as well as effete downmods via sockpuppets to *try* vainly "hide it" -> http://slashdot.org/comments.p... )... apk

Re:Totally this... what else could you possibly wa

[coofercat]

How about "the ease of distribution of Slackware 2.1" (ie. came on 73 floppy disks you had to download each one via FTP over ~64K lines) ?

Apart from in very rare cases, yes it does

[dbIII]

Well that is the very rare ideal situation of enough memory for the system in all cases and turning the machine off before it has cached much so I really don't get why you are mentioning it. What is your point exactly? That you've completely forgotten that cached stuff sometimes ends up in swap unless you take steps to make sure it does not?
If the system knows it has swap it will eventually use it unless you tell it not to.

Re:Apart from in very rare cases, yes it does

[Demonoid-Penguin]

Well that is the very rare ideal situation of enough memory for the system in all cases and turning the machine off before it has cached much so I really don't get why you are mentioning it.

Do you have something reliable to back up that statement? Seriously? It doesn't even make sense - how the fuck will a swap file/partition get written to (if it even needed to) if there's no power?!!

Your Windows "knowledge" does not translate. The default setup for modern Linux servers running on SSD (of which there are millions in production use) is no swap files or partitions. Power failure will result in identical data loss regardless of whether there is a swap file/partition or not. (even then there's journaled file systems)

Re:Apart from in very rare cases, yes it does

[dbIII]

Do you have something reliable to back up that statement?

I've been using the ganglia monitoring suite on clusters for a bit over a decade and on several hundred occasions I have noted that eventually cached stuff ends up in swap even if there's still some free memory. It's part of what it is for, getting stuff out of the way so you've plenty of memory to use.

how the fuck will a swap file/partition get written to ... if there's no power

WTF? Is this some sort of game where you NEED to win and are prepared to go beyond the far side of crazy to do so? If so just play with yourself and leave me out of it since I'm having a discussion not a game.

The default setup for modern Linux servers running on SSD (of which there are millions in production use) is no swap files or partitions.

It has always been an option and not bad for desktops but the default on servers - no that's just stupid since swap is often the difference between eventually getting something done or a complete crash leading to possibly hours of downtime.

Re:Apart from in very rare cases, yes it does

[Demonoid-Penguin]

Do you have something reliable to back up that statement?

I've been using the ganglia monitoring suite on clusters for a bit over a decade and on several hundred occasions I have noted that eventually cached stuff ends up in swap even if there's still some free memory. It's part of what it is for, getting stuff out of the way so you've plenty of memory to use.

how the fuck will a swap file/partition get written to ... if there's no power

WTF?

turning the machine off before it has cached much

You made the stupid claim - I just pointed out the obvious.

The default setup for modern Linux servers running on SSD (of which there are millions in production use) is no swap files or partitions.

It has always been an option and not bad for desktops but the default on servers - no that's just stupid since swap is often the difference between eventually getting something done or a complete crash leading to possibly hours of downtime.

I'll grant you that it's your opinion (which you are no doubt deserved of holding) - but it still doesn't change the reality that many million of enterprise grade VPS do not have swap. Doesn't mean your wrong. Likewise the bloke who lives on orange juice alone. In your case you trade performance for a few dollars saved on RAM - without knowing your use case I wouldn't call it stupid, but equally I'm wary of saying it's clever.

Re:Apart from in very rare cases, yes it does

[dbIII]

With respect, it's reducing the chance of outages at the cost of some possible slight lifetime loss on the SSD and has nothing at all to do with saving on RAM because sometimes things go wrong and whatever RAM you have is never enough. That's why whatever tech journalist who fed you the line about your mythical industry default is incorrect. Some people do it but the downsides are too great for it to be widespread.

Re:Apart from in very rare cases, yes it does

[Demonoid-Penguin]

That's why whatever tech journalist who fed you the line about your mythical industry default is incorrect.

I don't know where you got the idea I read it - it's my first-hand experience. (did you think my tech creds are playing video games?)

Not hard to test your assumptions - try googling for SSD VPS and then looking at what you get from various companies. The most common default in the sub-$20 pm range is no swap.

Re:Apart from in very rare cases, yes it does

[dbIII]

did you think my tech creds are playing video games?

Considering you wrote "Your Windows "knowledge" does not translate" to me, and your deliberate stupidity with the powered off thing how do you think it looks?
On topic again, back when SSDs had very short read-write lifetimes it was a bad idea to use them for swap. The world has moved on but some tech journalists etc got left behind, as probably did the bargain basement VPS deals too. Just because it's a cheap deal doesn't mean it's a good idea if you want stuff to stay up.

Re:Apart from in very rare cases, yes it does

[Demonoid-Penguin]

did you think my tech creds are playing video games?

[...]back when SSDs had very short read-write lifetimes it was a bad idea to use them for swap.

Agreed

The world has moved on but some tech journalists etc got left behind

I don't see the relevance.

, as probably did the bargain basement VPS deals too.

That's your opinion - we made it a policy because it made no sense for the business model. And sub-$10pm (AU) is bargain basement.)

Just because it's a cheap deal doesn't mean it's a good idea if you want stuff to stay up.

Agreed - but that has nothing to do with whether you are running a swap. And this is not "back on track". The track was "how do you enable permanent swap" - the answer is have a swap partition/file and set swappiness.

For more information please re-read.

Re:Apart from in very rare cases, yes it does

[dbIII]

the answer is have a swap partition/file and set swappiness.

Which is why I can't understand the bile of "Your Windows "knowledge"" and the deliberate stupidity about powered off systems when I made a few short comments right out of linux memory management for newbies. Did you get me mixed up with another poster or were you looking for a fight?
What was so deeply offensive about "If the system knows it has swap it will eventually use it unless you tell it not to"?

Re:Apart from in very rare cases, yes it does

[Demonoid-Penguin]

[sigh]

the answer is have a swap partition/file and set swappiness.

Which is why I can't understand the bile of "Your Windows "knowledge""

Your failure to understand and conflation of "I dinna calculate RAM demand" with "lucky I have swap" is your own problem. Have you heard of planning?

and the deliberate stupidity about powered off systems

I'm glad you (finally) got that. I don't hold it against you - but you really should attribute the stupid scenario to the person who suggested it (Hint: it were you).

when I made a few short comments right out of linux memory management for newbies.

There's your problem - right there. if you're going to talk trash like "bargain basement" (as if you are aspiring to ITIL standards) don't mix scenarios and quote from "n00b" guides (it's called "lugging the goal posts"/shifting grounds). The Google university isn't knowledge.

What was so deeply offensive about "If the system knows it has swap it will eventually use it unless you tell it not to"?

I'm not your mama (poor soul). If you insist on riding the bike without a saddle down bumpy hills don't come crying to me about butt hurt.

WTF is your issue? The original question (go on, look it up, the words and thread hasn't changed) was "how to enable permanent swap" . The answer was, and still is, set the correct level of swappiness for the swap (as others have pointed out, much more civilly - but you ignore that also to pursue a bad emotional investment). What the n00b guide at a cursory read didn't tell you (it's a guide for n00bs, not Enterprise BP with an SLA that bites), is that swap == performance loss (allocate more RAM, and adjust your plans))

Next time you launch your mouth attach less ego, more brain, and less ambiguity. Maybe write a business plan and include an escape clause.

You're mixing personalities with facts. If you make yourself a prisoner of the opinion of pseudonymous poster you're a dick. In this case, patently and demonstrably - you're both. Get over it and move on, hopefully the wiser. If you're posting to /. hoping to make a rep - you've got it wrong, on a number of levels.

It ain't news, it's opinions - some informative, be informed or waste your time. Pick one.

[...]were you looking for a fight?

Grow the fuck up. What are you, 16?

Re:Apart from in very rare cases, yes it does

[dbIII]

Your failure to understand and conflation of "I dinna calculate RAM demand" with "lucky I have swap" is your own problem. Have you heard of planning?

You really are pushing that "did you think my tech creds are playing video games?" line pretty hard aren't you?
Obviously swap is part of planning for when more RAM gets used than you have calculated - stop acting like you got all your tech skillz from playing video games and try wandering back to reality.

Grow the fuck up. What are you, 16?

So says the guy who led with the platform hate attack and followed with the weirdness about systems turned off. What is your game here? Why go rabid over little quotes from linux memory management for newbies?

Re:Apart from in very rare cases, yes it does

[dbIII]

WTF is your issue? The original question (go on, look it up, the words and thread hasn't changed) was "how to enable permanent swap" . The answer was, and still is, set the correct level of swappiness for the swap

"swapon" is the correct answer, yours is an answer to a different question (that was not asked) about tuning it.

Re:Apart from in very rare cases, yes it does

[Demonoid-Penguin]

WTF is your issue? The original question (go on, look it up, the words and thread hasn't changed) was "how to enable permanent swap" . The answer was, and still is, set the correct level of swappiness for the swap

"swapon" is the correct answer, yours is an answer to a different question (that was not asked) about tuning it.

[sigh] creating a swapfile and running swapon does not enable permanent swap. It only enable the possible use of swap - if the system runs out of RAM. Even pushing swappiness up doesn't make it permanent swap. Maybe you read all that n00bs guide. Surely it has a section about /etc/fstab.

Re:Apart from in very rare cases, yes it does

[dbIII]

It only enable the possible use of swap - if the system runs out of RAM

Obviously. That is the major reason to have it in the first place!
WTF is your problem and these "saddle down bumpy hills" streams of shit? WTF is your problem and the platform hate from nowhere when I don't even use the other platform? Why are you so intent on mixing personalities with facts and blaming me for your own stream of shit?

Re:Apart from in very rare cases, yes it does

[dbIII]

Grow the fuck up. What are you, 16?

When you were playing junior football in the late 90's I was running prac classes for engineering students (since I was taking a break after a few years in the power industry due to jobs being short) and doing a bit of consulting work before getting involved with using clusters to solve heat transfer problems then going full time into IT.
What is it with these stupid bullying bluffs? Besides, from your childish posts I was far more mature at 16 than you are now Peter Pan.

FTFY... apk

[Anonymous Coward]

Translation: I am mentally ill and need help, but the nature of my illness prevents me from realising this. I'm so ill, in fact, that I think it's a good idea to act like I am something I'm not in a fake it till you make it wannabe "Sigmund /. 'SiDeWaLk-ShRiNk'" fashion which shows my DELUSIONS OF GRANDEUR and can't be defeated by normal antipsychotic drugs I am supposed to take, and fail to see the irony. I will stalk and harass those who point out I am a cowardly little douchebag who is a hypocritical pot calling a kettle black weakling and mental health isn't something I ought to give advice on seeing as I am ill there too, to the point of alienating even more and more people, ensuring the cycle of my mental destruction will continue unabated!

Amiga Things I Wish I Could Get In Windows, OS X

[damnbunni]

Those are all things I do miss from the Amiga, but here are the things I most miss:

The focus paradigm. Windows and Mac use 'click to focus, focus raises window to top'.

Xmouse is 'hover mouse over window to focus, click raises'.

Amiga did 'click focuses but does not raise window'. It combines the best of both - you can focus a window that's behind another one, and you don't have to worry about tapping the mouse (or trackpad!) and changing the focus because the pointer moved. You could also double-click to raise the window, and windows had a depth widget that would raise it to the front or push it to the back of the stack.

The other thing I miss are DataTypes. The easiest way to explain DataTypes is 'Video codecs, but for everything.' Install a PNG DataType and any program that works with images suddenly understands PNGs. There are text format DataTypes, sound format, video format, etc.

um

[Anonymous Coward]

You're missing a lot. Here's just one thing you and lots of Linux developers are missing:

Try that again.... without a high-speed internet connection.

Modern Linux distros are becoming unusable outside of big cities. No longer can you grab a CD or DVD and go to an unconnected location and do a good install and setup all the applications - everything wants to get onto the 'net to download blobs and resolve dependencies... then everything wants to continually hit the 'net to check for updates. It's a mess that seems to be only getting worse.

If your OS install involves an internet connection, you are not in control, you have no stable reproducible config, and you have no security.

Re:um

[BevanFindlay]

This applies to most OSes now though. Windows needs a lot of drivers from the Net on any machine I've ever used, and has just as many updates - more on a fresh install. Mac OS X has stupidly-large updates on a fresh install (several GB). Also, I'm not a developer. :-) I don't know of any OS anywhere that is designed to work separately to an internet connection now, so I'm not really sure what your point is...

Re:For desktop OS, I'd tale BeOS' responsive handl

[Anonymous Coward]

perfectly responsive to user commands

I personally find that very important regarding window manager. If I command a window to be moved, resized, sent to the back or iconified, I want that to happen right away. No asking an application that is never going to respond for permission, leaving an unmovable greyed out zombie window in the foreground.

TempleOS

[bradgoodman]

...the religious rhetoric of TempleOS...

I would choose

[Anonymous Coward]

DirectFB, OSS and IOKit

VMS' fine-grained privilege system

[david.emery]

It was very easy to set up specific privileges for specific users/classes of users. I haven't seen anything else come close in 35 years in the business.

I also think VMS was the easiest system to administer, including a well thought out integrated help system that had the right answer for your question with minimal fuss.

Frankenstein or Dracula? You decide

[allquixotic]

My "build" of an OS out of constituent components would be:

  - Pure 64-bit; never never never never never any 32-bit support whatsoever throughout the software ecosystem
  - The Linux kernel
  - Solaris Zones (containers) able to host the latest Linux userspaces as well as an optional BSD and Solaris userspace with no virtualization
  - ZFS (okay, probably the latest version from Oracle is better than what Illumos has, so let's go with that)
  - An open-source version of Microsoft's WDDM as the graphics hardware abstraction layer (drivers are then built on that and are fully open source)
  - The best of Linux cgroups and namespaces reconciled with Solaris Zones
  - For a hypervisor (if you need to run Windows), Xen dom0/domU would be available
  - Dtrace from Solaris
  - kdbus
  - systemd core, but omitting many/most of the optional components (available as packages but not installed by default)
  - RPM for the package format, including Delta RPMs (drpms) for updates and LZMA compression on the package payloads
  - aptitude or yum for the package management interface / downloader
  - GNU bash
  - Entire system compiled with clang by default, but with gcc available as a working alternative (competition is good; One Compiler To Rule Them All is bad for progress)
  - A fully working, optimized, functionally validated Win32 and Win64 emulator (including graphics libs) supporting Windows Desktop apps that require any version of Windows from 95 to 7, for those legacy apps that just won't die
  - Both the latest open source versions of Java and .NET installed, available by default, and automatically updated with no nags, but with neither one shipping any browser plugins
  - No Flash!

Windows, Linux, Solaris, AS/400

[fluxsmith]

Windows GUI, Linux kernel and CLI, ZFS filesystem, AS/400 storage addressability.

I wouldnt

[Anonymous Coward]

If I had the knowledge and time, I'd build an OS from the ground up! No registry, no config files, no command line or terminal windows. The basic OS would run on two files. System (the kernel) and Finder (the GUI) Then from there to add functionally to the OS, plug-ins (drivers) would be packaged into 1 file and dropped into Extensions folder. No complication, no clutter.

Fore those who haven't figured it out yet, I'm refereeing to the old Mac OS. :-)

Still my favorite OS of all time, LONG LIVE SYSTEM 7!!!!

The VMS Common Language Environment

[stevel]

I'll admit that I am biased, as a former VMS developer for DEC, but in my opinion VMS did one thing right from the start that I have not seen any other OS duplicate before or since - the Common Language Environment. VMS defined a common calling and exception handling standard that was used by all of the 20+ programming languages supported on VMS. The system services and the common run-time library were usable from all of the languages. Yes, many of the languages needed extensions to support things such as "pass by descriptor", but it was done in a consistent fashion. There was also a naming standard that separated system and user namespaces to avoid namespace collisions. This was all documented in the standard VMS manuals and was designed to be extended as needed.

This also meant that pretty much all of the system library routines were language-independent and there were large collections of these that could be called from most languages. For a long time, Windows had something close to this with the Windows API, but in recent years it's been shifting to C++ class libraries that shut out other languages.

Re:The VMS Common Language Environment

[lophophore]

Free advice is worth every cent, Steve. Wasn't that you?

Re:The VMS Common Language Environment

[stevel]

Free advice is worth every cent, Steve. Wasn't that you?

Yup.

I should also have mentioned that the common language environment meant that mixed-language applications were far easier than on most other operating systems. How about mixing BASIC, RPG II, Pascal, Fortran and Ada? Easy.

Tandem Non Stop Kernel

[Anonymous Coward]

Tandem NSK and hardware, whatever you want to run on top is fine.

This is a question with multiple faces.

[Anonymous Coward]

"If You Could Assemble a "FrankenOS" What Parts Would You Use?"
Let's throw some cognitive review about the question:
This is a number of questions rolled into one, for a start, you would need to identify who you're asking:
YOU:
The non-tech user
The media-consumer user,
The "online" user,
The tipical internet user,
The power user,
The developer,
The Admin,
The Gamer,
The Hardware guy
As this of each users have different needs, and would be better served by a quite different OS, in real life, their requirements have evolved into four OS's: The smartphone user, The general user, The power user and the server-admin user, where each individual user will take the pros and cons of the OS environment and then set it up so it better works for them.
IF:
Here we're defining really who we're asking from, this question is dedicated therefore to only some of the initial users: the power user, the developer, the admin or the gamer, as this kind of users will be the only ones able to -not only understand, but compare the different OS's mentioned and experienced, and identify where each of this one fits into an overall experience, so far it's fine, the niche of /. is this same market, most of the people here should be able to understand and reply to this question, but it still insures at least 4 different responses to the original question.
FRANKENOS:
Hipotetical in nature, this word allows us to forget what we have learnt in the way, and imagine what are the best parts of each OS we have worked with, but it doesn't consider the different moments each of us have been introduced and used each OS, it not the same to think about Linux in the pre-version1 that in current times, as it's not true about Windows, OS/2 or MacOSX, not to mention the variations in the hardware each of the OS's runs. Honestly, this makes any potential response invalid, but then again, making it totally unreal, it allows us to dream, but as they say, "one person's dream might be another person nightmare"
I say it.
So, what would i reply? I loved OS/2 at the time, i have always hated Windows with a passion, and i use MacOSX as the currently best, but there is no justifiable answer to the original question, and i don't like open ended questions, please improve.
LA

Re:For desktop OS, I'd tale BeOS' responsive handl

[tigersha]

I still have an old ThinkPad with 256MB Ram that I have running on BeOS in a drawer.

BeOS is very responsive because EVERY thing ins threaded and they embraced multi-threading and SMP from day 1. Other OS/s use an event loop in the GUI apps while BeOS does it somewhat different. It encourages threading and not a single loop when you consume an event, process it and then go on to the next one. Events tend to spawn threads that then interact with each other. The single event consumer thing is a very old paradigm in UI design.

Another OS which is very very responsive for the same reason is Photon, the GUI running on QNX.

I miss BeOS, was really a great system.

VMS on x86_64 with full POSIX

[lophophore]

So I'm not biased or anything. But VMS (real VMS) with the full POSIX support, X Window Systen, running on commodity hardware with source, well, that would be pretty cool.

If I have to make a Franken-OS, I'd like to take the concept of logical names from VMS and bake that in.

For those of you that don't know what that is, you DEFINE a name, to another name, or list of names. You can use that new name like any other name in the system, including the name of a file system. This seems useless on the surface, but the real power is that you can define a name to a list. This could be compared to your unix shell PATH on steroids that have been taking steroids. It is one of the key concepts that lets VAXcluster work. It makes a file system location work like a search path.

I'm a little rusty (is has been almost 25 years) but as I recall, on a cluster node, the logical name SYS$SYSROOT mapped to a list of SYS$SPECIFIC, SYS$COMMON, which allowed the system to have local configuration files that overlayed the cluster's common files. You could set up a logical name for anything that could be named, devices, queues, users, whatever. Very powerful.

memory safe os, no mmu

[Anonymous Coward]

I know I sound like a heretic because I have learned to hate Unix and c. And all it's likeness such as NT and osx.

The whole model is fucking broken. A single bug will enable attackers to take over EVERYTHING. No such thing as fault isolation.

Memory safe languages can offer us a much safer future. We would not even need an mmu and no hard distinction of kernel and application.

Apparently Burroughs, elbrus and iCl had this kind of thing a long time ago.

Now we all eat the bell labs fast food dreck.

Re: memory safe os, no mmu

[Anonymous Coward]

We already have swift and rust, now somebody needs to write an os with it.

And yeah, call it franksOS. That would be fitting given the amount of work done by a certain Frank.

Re: memory safe os, no mmu

[Anonymous Coward]

While we are at it, teach Computer Archeology. Too many insane regressive ideas are currently in need to be shitcanned.

Nsa and gchq are the modern day Francis drakes. Terrorizing business in the name of myopic interests. Also see the wassenaar nonsense the pulled recently.

Tim cook and Sergey brin must reign into these Muppets.

Re: Amiga Things I Wish I Could Get In Windows, OS

[Anonymous Coward]

I'm not convinced by the hover to focus idea in any guise. I'm of the "choose the window you're working in, get the mouse the hell out of the way and work" school. Getting the mouse the hell out of the way usually involves randomly shoving it to the side/corner of the screen. If there is some sort of hover to focus thing going on and a window happens to be there (or worse yet some version of "put mouse to side to gets me m'charms") that really screws with my preferred work pattern.

Re:For desktop OS, I'd tale BeOS' responsive handl

[Anonymous Coward]

BeOS didn't even have the concept of a wait cursor, because the system was never in a state where the user couldn't use it.

Re:For desktop OS, I'd tale BeOS' responsive handl

[Anonymous Coward]

In addition to your addition, BFS allowed completely arbitrary metadata for any file. These days we are finally seeing file systems that allow tagging and whatnot, but BeOS allowed you to tack anything on to any file. Want to embed a description? No problem. Want to embed a multi-gig video file? Still not a problem.

Re:Try and make an OS that viruses couldn't target

[cbhacking]

Or just hook the keystroke window messages in the victim apps. You can do that using the debug APIs (assuming you are executing, and the other process isn't more trusted than your process or in a different user session), or by setting Image File Execution Options (requires Admin) to tell Windows to load a specific DLL into every process...

Re:For desktop OS, I'd tale BeOS' responsive handl

[daedalus2097]

BeOS is excellent in that respect, and similar to Amiga OS in that and many other ways too. It's something I still miss, being able to click a button and see it react, knowing that even though my computer's 100% busy doing something, it still has time to acknowledge my request. Another aspect I would keep from Amiga OS is the ability to still manipulate an application's windows, even when it has a requester (dialog box in Windows parlance) open. So if some application is looking for attention, I can move, resize, send to the back or minimise its window and requester to get on with other work until I'm ready to deal with the open requester. The concept of screens (as opposed to multiple desktops as is standard on Linux distros) from Amiga OS, and the REXX messaging and scripting system is something that no other OS has done so nicely or comprehensively.

So, mostly Amiga OS things for my dream OS then ;)

Re:Totally this... what else could you possibly wa

[BevanFindlay]

Ooh, forgot that one. My bad. And the browser would have to be one of the awesome 1.0 versions (of pretty much anything), and doesn't get fixed for years (so pretty much IE). Or maybe Linx, because who needs images or interactive content or anything?

Re:Totally this... what else could you possibly wa

[BevanFindlay]

Yessss... And the file access capabilities of a non-jailbroken iPhone. :-) Thank you for getting me out the M$-only slump I was in! (Of course, I've heard that you could get Windows 95 on floppy disks, and it wasn't apparently much better than the Slackware example).

File versioning first used in IBM OS/360 in 1966

[Anonymous Coward]

IBM today may be totally dysfunctional and likely doomed to collapse in the next few years but back in the 1960's IBM invented most of the Operating System and CPU Hardware ideas we take for granted today [Mufti-processors, Multi-threading, Task Schedulers, Virtual Memory, Virtual Machines].

The first version of IBM's mainframe Operating System, OS/360 MFT had versioned files called generation data groups. You could specify the maximum number of file versions and the expiration dates or time periods where old versions would be automatically deleted.

If you want to prove you are a serious hard-core geek; the final version of OS/360 is in the public domain and can be freely downloaded and run as a VM using the Open Source Hercules emulator.

I can't rexactly say what parts from other OS's..

[Druegan]

... since I don't really know all that much about them. (Yes, my nerd-fu is weak.) I can, however, give some features and behavioral descriptions of what I'd like to see in a "FrankenOS".

"Userspace Stuff:"

1.,"It just works". Really good hardware detection and a set of generic drivers that will at least grant basic functionality for all essential hardware.

2. Robust documentation and error messages. When something breaks, I want to know what happened in real language, even if it's just an approximation, not some string of gobbledygook that I need a degree in CS to begin to parse. I also want clear documentation on what every part of the OS *is*, and does. Configuration options, running processes, services, threads.. the whole bit.

3. Windows level gaming capacity. Dx-whatever-the-latest-and-greatest-is.

4. A "no-bullsh*t" UI. I don't want any fancy-schmancy crap like transparent windows and animations, or retarded nonsense that tries to make my desktop look and act like a phone. No stupid behaviors like if I drag a window up to the top of the screen, it changes the size/orientation of the window.. none of that. Just let me set a picture as a background, and give me good configuration options so that other things could be added in as "plugins" if I should ever, for some godforsaken reason, decide I want them. Just make them easily unpluggable for when I come to my senses. Multiple desktops are reasonable.

"Under the Hood:"

1. Native multi-level sandboxing with encryption. "The issue isn't whether you're paranoid, but whether you're paranoid enough." Nothing that accesses the internet should get within 2 layers of the core OS. Options should exist to lie to the "internet layer" about every aspect of what the system is, does, is running, etc. from window size to OS version. Nothing accessible by a browser should remotely be able to see anything outside the box the browser is contianed in, inclduing memory spaces and the like. Ability to handle encrypting everything from file systems to memory to the bootloader. I don't care if it slows things down a bit. I want "paranoid about privacy" function.

2. Modularity by design. Components should be compartmentalized, and as independent as possible. I should be able to strip out, or not include, any element of an install that isn't explicitly necessary for the funtionality I require. Do one thing, and do it well, but that's it.

3. The ability to set hard memory limits on a per-application basis, and a really nice memory management system that polices those looking for crud and cleaning.

4. Comprehensive versioning management relative to the OS itself, and the ability to "roll back" updates that make a mess due to being broken. Similarly, an in-built VM that lets you test out any prospective updates in its own sandbox prior to them being allowed to go live.

5. The ability to "hard kill and purge" any app that is found to be misbehaving. As in "immediately brute force stop it from doing what it's trying to do, and completely flush the memory space." Preferably by some keystroke combination that doesn't involve needing to have some kind of task manager always running to execute.

6. A good, intuitive, logical/physical drive management system for all OS functionality and components. Specify what drives/segments are used for OS files, swaps, temporary files, sandboxes, etc.. each with versioning control, journalling, etc.

7. Specific update management sorting. Deliberate blocking of "schedulers" and "auto-update tools" and the like.. permissable only maybe in certain areas of sandbox nests. It's manual and requries explict user enaction, or it doesn't happen, period. Same with "helpers" or any other of these bullsh*t "always on in the background" programs that companies love to infest systems with these days.

"Ideological Stuff:"

1. "Free as in Freedom, not necessarily as in Beer. But beer is good too if it can be swung."

2. "F*ck DRM"

3. "F*ck advertising and tracking a

Re:For desktop OS, I'd tale BeOS' responsive handl

[blind biker]

BeOS didn't even have the concept of a wait cursor, because the system was never in a state where the user couldn't use it.

Indeed. The user was King - at all times and in all circumstances.

Don't blame me for voices in your head

[dbIII]

but you really should attribute the stupid scenario to the person who suggested it (Hint: it were you).

Quote it to me.

Maybe I have to be even clearer

[dbIII]

It only enable the possible use of swap - if the system runs out of RAM

Yes it gets used when the system thinks it needs it and not when it doesn't. So never come close to running out of memory and don't have the system up for months caching stuff and eventually dumping some of it into swap and it will never get used at all. I've go no idea why you want to argue against the obvious that we both agree on.
Is that dumbed down enough or do I get another pile of deranged hillbilly shit thrown in my direction because you've been having a bad day or something? WTF is your problem and why are you taking it out on me?

What about the compile-your-own distros ?

[RockDoctor]

Never tried one myself, but unless I've deeply misunderstood things, they allow you to build your own choice of modules into your own design of system?

For QC of the build process, I'd expect the purveyors of "build your own distro" distros to ask to know what they need to play with (not everyone will answer. [SHRUG]). so, surely a first step would be to look at their buld records and feature request lists.