A Praise To Unix

MotyaKatz writes: "ZDnet has an article from Evan Leibovitch which he calls The Unix Phoenix. As he states, 'I come not to bury Unix, but to praise it'. He mostly deals with the aspects of Unix surviving during Linux growth."

Comments

Re:Article was fluff

[benedict]

On the other hand, I am told it is much easier to become a Linux developer, so Linux is more open.

On the other other hand, there are fewer license restrictions on what I can do with BSD software, so BSD is more open.

On the other other other hand, Linux's license promotes open-source software, so Linux is more open.

On the other other other other hand, "open" is a word with a lot of meanings and a lot of connotations, and perhaps there are better things to do than worry about which project is "most open".

--

The problem with today's users

[Fervent]

The problem with today's users is that UNIX-mavens naturally assume what they like and use is what the general populace should use. The truth is, regular users don't really care about the same features that mavens like.

Case point, I tried installing Linux on my family's home machine. I tried to explain that the system would rarely (if ever) crash, that each person in the family would get their own desktop, and they would get the access to the internet they always had.

You know what? They hated it.

'Why do I need to type a password every time I want to get on a machine I own, in my home?' my mother complained. The kids said they couldn't install any games they liked, and the ones they could rarely ran. They wondered why they needed something called a 'superuser' to install Q3A.

And here's the kicker: 'Why do we have to worry about crashes?' To them a crash was standard behavior, like a car misfiring on startup or a TV channel fizzed out for a few minutes. 'But then you won't have to reboot,' I'd explain. 'But we have to reboot anyway. We shut off the machine every night.'

You see, the practices of the UNIX/Linux world don't seem to jive with the world of "Joe User". They've ascribed to practices that, for better or for worse, they don't want to deliniate from. Case in point: the whole VHS vs. Beta debate a while back. Not looking at which format was "better", once people started using VHS they never looked back.

Same thing with Windoze PCs. Once people started looking at the pretty icons, the ability to run all those games, relatively simply document/folder analogy (even though copied from Mac) and the ability to use all kinds of neat hardware, they were hooked. Password protection, uncrashable system? This was a family computer living in the den. They didn't care what was the pride of sys admins, web servers and academic folk in college. Unix never existed in their mind, and if it did, it was that brief glimmer of "nerdiness" they all wanted to avoid.

My commentary on the state of technology and the world.

Re:Linux rehashs 70s era OS.. wow, special.

[WhyteRabbyt]

I think the real future of Unix looks something like MacOS X, not Linux. Only if you think Unix becomes a single-vendor prospect. Building a (admittedly very nice, very clever) GUI on top of a Unix core however is not appropriate for all the areas Unix is used in. Desktop/Workstation areas, yes I think the Apple stuff will have an influence, but I dont Apple are about to start competing with the heavy-duty kit SGI, IBM and Sun produce. I dont see Apple's new GUI as being particularly relevant to servers. And thats Unix primt-time.

Gnome and KDE are just the first iteration towards a useful user experience.

Yup, just first iterations. How many iterations of MacOS did Apple do, before they dumped most of it for MacOS X? How many iterations of Windows have there been? I get the impression both KDE and GNOME are progressing far faster than either of these did.

Linux so far is a step sideways at best.

Sideways from what? A GUI is not an OS. You might think that an OS without a fancy GUI isnt useful, but that doesnt make it a fact.

Pax,

White Rabbit +++ Divide by Cucumber Error ++

Re:JWZ

[WhyteRabbyt]

said it best: "Linux is only free if your time has no value."

Whereas with Windows NT you're stuffed both ways...

Pax,

White Rabbit +++ Divide by Cucumber Error ++

The differences between *nix & Win*

[Darth RadaR]

/* just my opinion & experience. Your's may vary. */

Let's just face it, Unix's death has been predicted for years. Kinda like rock & roll music, it won't go away. Like rock & roll, Unix always evolves and pops up with different flavors and tools to suit the times & needs.

Unix is not for everyone. I agree that it's a bit much for someone who just wants to send mail, play games, use AOL, and do word processing. But for people who really need more control & precision over their systems, Unix is the way to go.

It took until Win 2K server for M$ to realize that file quotas might be a good idea in their OS, whilst Unix has had that for years. That's how Unix always evolves around realistic and current needs more immediately, whilst (IMHO) Win* is either catching up, or putting out vaporware (i.e. J++).

Comments on replies

[Animats]

what he suggests is far beyond squeak (i've never used genera, so I can't say). He wants efficient, secure cross process objects.
Yes. Things like this has been tried at the language level. Java is the most visible example. I'm proposing to do it at the OS level, where the protection hardware can make many of the checks. That way you're not locked to a specific language, and can reuse more existing software without trusting it.

So what do you think about the Spring and EROS papers? You have read the Spring and EROS papers, right?
Spring was an interesting idea. Many of the ideas ended up in Java, but once they started thinking cross-platform, they were stuck with running on top of an OS, not writing one. Yes, there was a JavaOS. Haven't heard much about it lately.
As for EROS, I was impressed with that project, even though it never quite got finished. They had several problems, some technical, some political. They never really explained how to get from capabilities to policy. Capabilities are a low-level mechanism, and it's not clear how you get to, say, a web server with secure server-side applets, or its dual, a web-browser with secure client-side plug-ins. That's a classic failing of the capability crowd, all the way back to Norm Hardy and KeyKos. Too much abstraction and not enough explaination. I also tend to think they overemphasized persistence. The opposite of persistence, the pure transaction system, is closer to what people want today. Consider cgi-bin programs, where the transaction program is flushed at the end of each transaction. Hokey schemes like running Perl in the web server's address space are needed to make this efficient. That's a concrete example of where the OS has totally inadequate mechanisms. The persistent part of a server belongs in a database where you have a coherent model of the data. Think about how hard it will be to clean junk out of a persistent object store. It creates the versioning problem from hell.

Congratulations, you've just described the foundations of Windows NT.
Early NT was closer to this than NT is now. "Event pairs" were getting close to what I'm talking about here. However, rather than fixing NT's IPC performance problem, Microsoft chose to kick Dave Cutler out of the top NT job and dump much of Windows 95 into the NT kernel. Now it's just another big, monolithic OS vulnerable to a bug in any of millions of lines of code. (My favorite discovery was that the decompressor for RLE-compressed .BMP files is in the kernel, and contains a buffer overflow.)

Strange, most of what you describe is MULTICS...
Not quite. Multics security is hierarchical. I'm discussing peer-to-peer, which is more useful. Multics did a good job at security, and was very well regarded in the DoD world. NSA's public access machine was a Multics until just a few years ago.

So you want a microkernel. Mach provides much of what you ask for, except the fine-grained security perhaps.
Mach offers a classic I/O type IPC primitive. Actually, QNX does a better job at that.

You are paranoid. You don't need an operating system; you need a psychiatrist.
What I want is security that actually works, rather than having to be patched every week. A system with tiny amounts of trusted code, in which untrusted code can't break security. Something where we don't have a major nationwide crisis every time some script kiddie tries something.

Re:Beyond the UNIX model

[Kaufmann]

Congratulations, you've just reinvented Genera. Or Squeak, if you like "objects" better than "functions", or if you want to run it on top of an existing OS, or if you don't want to be tied to a specific (dead) hardware architecture. And ETH Oberon is yet another OS based on the same ideas. In any case, that you don't see it everywhere doesn't mean it hasn't been invented yet - and that it's (not) popular doesn't mean it's (not) good. (For the first case, see Haskell - "groundbreaking" parametric polymorphism in the late 80s; for the second case, see Windows.)

In any case, it will do you no good to use CORBA as it is today. Instead, use a dynamic, high-level language for user-level functionality, and just let applications people deal with objects in the language's natural idiom, making no syntactic distinction between "local" and "remote" objects.

In any case, have fun, and don't let those Unix weenies tell you that systems research is dead - if it were for the conformists and the naysayers, we'd be rendering polygons with abaci!

Re:Linux rehashs 70s era OS.. wow, special.

[Black Parrot]

> Linux so far is a step sideways at best.

Maybe so, if you compare Linux to Unix as a contextless technology.

However, if you look at the role Linux plays on the desktops of those of us who use it there, Linux is a huge step forward over what most of us would be using otherwise.

--

They just don't get it.

[WiseAcre]

people who talk about linux and unix as a user-platform, comparing it to the Windows user experience or the Mac user experience just don't get it: Unix and now especially linux are programmers' operating systems, designed and built for programmers. Programmers who enjoy coding so much that they'll do it for free like to use unix, and not Windows.

People who poo-poo the potential that linux has need to remember two things:

• in the future, they'll get what the programmers give them. that's the nature of software. if the programmers choose linux (for handhelds, etc.), then that's what users will be using.
• don't believe me? in the past, they never predicted where linux would be today, so that's why we're not listening. and I, as my friends will tell you, was predicting exactly what has happened, and I was predicting it five years ago.

Re:They just don't get it.

[Elvis Maximus]

programmers are secondary to users!... Userbase is king. Windows has the userbase. If Linux doesn't get a big userbase, then people won't develop there.

That's true if the developers are selling what they're developing. If they're giving it away for free... well it kind of turns that model on its head, doesn't it?

-

Re:They just don't get it.

[be-fan]

Aside from the fact that that's awefully arrogant of you (programmers are secondary to users!) you're theory doesn't hold water. Two points.

A) Userbase is king. Windows has the userbase. If Linux doesn't get a big userbase, then people won't develop there.

B) Developing for UNIX is not necessarily better.
- There are a lot of commercial developers that grew up using Microsoft code generators. Those people will be the last to switch to *NIX.
- For a lot of smaller developers (database developers, etc) the fastest way to do something is still use generation tools (like database generators, or stuff like Delphi.) Those developer's won't switch until a similar body of apps are available on Linux.
- Any tool you can use in *NIX, you can use in NT. (GCC, Emacs, etc.)
- The Windows APIs are generally less chaotic. While Linux is in a constant state of flux, the Windows APIs are more stable. For example, in Linux, you have the major DE's changing heavily quite often. In Windows, the major DE APIs haven't changed all that much since Win95.
- Linux suffers from API overload. Which sound system do you program for? If you want featues and speed, you use ALSA. However, can you expect your users to have it? You could use OSS, but why not just forget it and use DirectSound? Which DE should you program for? KDE or GNOME? In the end, most commercial developers will just give up trying to choose and use Motif. And by using motif, they lose out on the cool features of KDE and GNOME. What toolkits? How can I make sure the user has the correct toolkit? What version of glibc? What version of X? You don't have these kinds of problems with Windows, mainly because MS forces people to stay on an upgrade path, and thus most users will have NT4, service pack 6a. And if they don't it's relativly easy to just include it on your CD, so you can be sure of the state of the system.
- In terms of mutlimedia, the Windows APIs are still much better.
- Developers still care about user experiance. When Sierra didn't let people choose the install directory in it's utilities, people had fits, and threatened to boycott their programs. In Linux, many programs have pages of installation instructions. Recently, I read a review of the 3 IDE's available for Linux. In the case of RedHat's GNUPro toolkit, the installation instructions consisted of half a page of CLI commands. Compare this to Visual Studio's, where you pop the CD, pick which options you want, and then go away while it installs.

Re:Missing Features?

[be-fan]

A lot:
Solaris: Massive SMP scaling.
IRIX: Mature 3D framework.
IRIX: Stable journeling file system.
Solaris: Dynamic patching of most kernel code.
QNX: Real time sheduling.
Solaris, NeXT, etc. Display Postscript.
NeXT, Solaris: Flexible, ObjectiveC object model.
And of course, most commercial UNIX's offer management tools that are much more integrated and functional that Linux ones.

Beyond the UNIX model

[Animats]

As an old-line secure operating system designer, I've been thinking about where to go after Linux. My current thinking is along the following lines:

• The main job of the OS is to facilitate protected inter-object communication. What's needed is the ability to make a CORBA-type call almost as fast as an ordinary subroutine call. Applications can then be built as lots of little objects in different protection domains. (For example, anything that looks like executable web content needs to be sandboxed.)
• Processes are not a primitive. The basic primitives are threads and address spaces. Threads can cross address space boundaries through call gates. This replaces inter-process communication.
  Why? Because when one object calls another, what you want is a subroutine call. Most OSs make you marshall the parameters into a buffer, make an IPC call that works like an I/O operation with one or more process switches, and then unmarshall on the receiving side. On return, these steps are repeated. All this is time-consuming. This is the main reason software isn't usually constructed out of little objects in different address spaces - it's too slow. Fix that problem, and clean design becomes efficient.
  How? The MMU and protection hardware in x86 machines has lots of stuff that's almost never used, like call gates and variable-length segments. If used properly, cross-address-space calls can be quite efficient. You don't have to copy everything, and you don't have to give up security to get performance. (Or so the data books indicate. I'm assuming all that stuff, like call gates and rings of protection, actually works.)
• With fast, secure object calls, the pressure to dump stuff into the kernel disappears. File systems, networking, and drivers move outside the kernel.
• Security is applied at the object call setup level. Look at the CORBA security model for a starting point. Security decisions are made during object loading and call gate setup. Once it's been established that object A can call method M of object B, no further OS intervention is required for that call.
• Very little code is trusted. Writing a mail handler? Just about the only trusted part is the object that handles putting the mail in a local mailbox. A web browser? Maybe bookmark handling and some certificate functions; everything else is sandboxed. Most code needs no more privileges than an untrusted Java applet.

This is an idea I've been kicking around for a while. With the GNOME crowd going CORBA, this is starting to look more practical. An OS like this will have something to run on it.

Comments?

Premise of article is utterly false

[Morgaine]

But Unix still has value that the Linux crowd may vastly underestimate in its haste to issue a death certificate.

What a total waste of electrons, both in the alleged view of the "Linux crowd" and consequently also in the article.

Repeat after me, 10 billion times: "Linux is a Unix".

Who cares a damn about the legal niceties (more like idiocies) that prevent one from using the label "Unix" where it's obviously appropriate. Ask any person with more than a little experience of Unix and you'll always get the same answer: Linux and the BSDs are all Unixes, through and through, every bit as much as the licensed proprietary versions. It's not just by accident either, it's by design. And in many ways (but not all, yet) they're the best Unixes around, with the older "legal" Unixes fighting hard to keep up. Anyone that thinks that the important thing about a "Unix" is its license is just so uninformed that it's sad.

OK, I know it's summer and good news is hard to come by, but that article was about as empty of point and content as they come.

Linux rehashs 70s era OS.. wow, special.

[weezel]

I think the real future of Unix looks something like MacOS X, not Linux.

The communinity development model so far has been unable to do anything other than kludge together something as important of the GUI. Gnome and KDE are just the first iteration towards a useful user experience.

Apple, on the other hand, has taken the core of a Unix system and used a single vision/goal/thingy to synthesize something new and exciting from two fairly stagnant OSs. Borrowing from the low level functionality of Unix and the elegant UI of MacOS they have made a real step forward.

Linux so far is a step sideways at best.

ahhh yes..

[fluxrad]

i long for the days of yore...with Bell Labs and the like. I long for the openness of the original unix. so i wrote a little play:

curtain. we see a developer somewhere in california

Developer: Boy, i sure like this AT&T Unix. Good thing it's open. I think i'll use some of the source for something

AT&T: Do that and we'll sue your bitch ass!

curtain

The original unix was free as in 'not free'


FluX
After 16 years, MTV has finally completed its deevolution into the shiny things network

Re:It's sorta strange...

[Octorian]

This is what most people think...

What everyone fails to realize, is that those systems continued to exist throughout the era of the PC. It's amazing how many people don't know that the PC has "not" been the be-all/end-all of computing in the past 10-15 years. Heck, until 4-5 years ago, the PC was an utter piece of crap compared to your average UNIX machines.

Look at some of what I've been using lately and the dates of manufacture:
Sun SparcStation IPX '92
IBM RS/6000 POWERstation 350 '92
SGI Indigo2 Impact '96

These WERE within the "era of the PC", are absolutely not PCs, and run UNIX in all it's multi-user, networked glory.

We never moved from high-end/UNIX to PC/Mac and then back. Both worlds existed in parallel, with very little communication between the two.

Re:It's sorta strange...

[Legolas-Greenleaf]

hehe... i'm quite aware that the large servers kept going with UNIX/VAX/etc. However, once the pc appeared, the average user didn't have to log into the mainframe to wordprocess.

we have an sgi iris file from 1992 at our work. despite it's age, it's a pretty impressive piece of equipment.

My point was simply this - the focus of the average user went from large timesharing systems to the personal computer, and now it's almost as if networked, timesharing systems have come back into vouge. i know, the whole while, there were multiuser computers running networks and databases and whatnot, but now we're even getting that sort of os on our desktop (i forgot to mention winnt as a multiuser os.)

it was just a dumb thought provoking-type comment. don't worry too much about it. =^)
-legolas

i've looked at love from both sides now. from win and lose, and still somehow...

Re:But is linux any good?

[WhyteRabbyt]

Linux is a mess of different libraries

I have 1067 .DLL files on my NT systems main drive. I'll bet you a shiny new dollar thats a hell of a lot more than any Linux system.

I am back to using Windows98SE and AmigaOS - why? One runs everything, and is faster for desktop use than *nix. The other is the only OS that had an advanced design and implementation. I really hate those fuckers who found out Unix in the past year or so and think it's fucking brilliant.

Maybe Win98 is faster, maybe it isn't. I work alongside someone who does OpenGL development, and he find Linux far faster running and compiling his code than Windows is, on the same code. So 'is faster' probably isn't true for everything. Linux is more stable. Thats a win for some people, more than speed. Meanhwile, BeOS probably has AmigaOS beaten in terms of implemented advanced design and implementation. Dont get me wrong, I loved my Amiga, and the OS was fantastic, but just because its the newest shiniest, doesnt make it better. I guess you're one of those fuckers who found out the new AmigaOS in the past year or so and think its fucking brilliant, eh?

Linux is the past. It's merely a free implementation of something thats been around for years.

And thats a problem why? Because its not new, it would appear. Sorry, but thats not good enough. Linux provides a solution for a problem. A modern Unix, capable of using a very decent subset of modern hardware, that has an aggressive development policy. Why is that somehow 'not good'? Lack of GUI apps? Sorry, GUI software doesnt make an OS better or worse, it just means you have GUI software. Desktop Linux is in its infancy. So what. Its progressing faster than the Mac or Windows ever did. And meanwhile, the core OS is still exactly what Unix always was; stable, secure, powerful.

Pax,

White Rabbit +++ Divide by Cucumber Error ++