Last I checked, Mandriva was proprietary. It may have a free version, but it depends on proprietary software. Right?
Ubuntu, by default, comes with entirely free software. The proprietary stuff is easy to install, but it's nice to know that most of the core stuff is free, and will always be free.
I tried to install Hardy Heron on a fairly stock, high-end Dell desktop earlier today.
Dell sells fairly stock, high-end desktops with Ubuntu preloaded. When buying a new computer, at least, you can let Dell go through all the pain of installing and getting it to work with your hardware -- and when it doesn't, you get a support contract.
Which is, by the way, exactly what you get from Dell when you buy Windows.
1) It was unable to use my RAID controller to install onto the existing RAID-1 array. It insisted on being installed on a single drive. To get it to boot at all, I had to completely break the RAID volume.
Linux does support software RAID, and many RAID controllers. But realize that by using hardware RAID, you are well beyond "the masses", and should be prepared to look this stuff up yourself.
Also, the server and alternate install CDs are much more likely to work with RAID and similar things, just as the desktop install CDs are much more likely to setup X correctly. If you know what you're doing, though, you can simply install from the server CD and pull in the missing components from the desktop.
2) It was unable to use my wireless card. It didn't see it at all, so it wasn't in the "connections" menu.
What brand is it?
This is a place where support is spotty lately. That's not an excuse, and not a way of saying it's "ready", but if you are interested, it should be pretty simple to find a list of supported hardware, including which wireless cards work the best. Anything Intel is probably going to work out of the box, with open drivers.
Unix for the masses is here, and it's called OS X.
And for the slightly less wealthy masses, it's called the EEE PC.
If they try to install Windows on their Ubuntu system, Windows will destroy their access to their Ubuntu partition with no questions asked.
Last time I installed Windows, it always warned when formatting. That is: It won't delete a partition without requiring a series of steps which are clearly designed to avoid the user just hitting next-next-next, yes-yes-yes, or reflexively doing whatever they think they need to do to procede.
The same is true of formatting -- the XP installer, at least, clearly warns that you won't be able to recover.
Now, of course, Ubuntu makes it incredibly easy to shrink your existing Windows install, whereas Windows provides no tools for dealing with other OSes at all (or that I know of). But it does warn you every step of the way, in typical Windows fashion, that you had better be damned-fucking-sure before you nuke any data.
Leave Linux to the power users and the server market.
As a power user, I would love for Linux to be mainstream. The more mainstream it gets, the more likely my video drivers are to work, and the more likely I am to have some decent games to play.
As a server administrator, I would love it if all of our developers ran Linux on their desktops. It's still possible to run into surprises deploying from Windows on their workstations (read: laptops) to Linux on the server.
Quite frankly, I don't want to use the same operating system as someone who refuses to edit any configuration file.
Here's the cool part: It's not up to you.
The thing is, Linux -- or, more generally, all open source software -- is for everything and everyone. If there's anyone who can't use it, or anything it can't yet do, that's just another problem to be fixed by anyone who has the time.
And no one can stop it. You can't make it into your 31337 high-school h4x0r club anymore. It's much bigger than that, now.
What stops another company from springing up to provide cable internet services for cheaper? Answer - government intervention.
Whenever I hear this, I always ask: Are you seriously suggesting that there be more than one company in a given area running physical cables to every house? Or are you suggesting more government regulation to force them to share the cables they've got?
Saying that government regulation is somehow going to fix what government regulation broke is absurd.
It sounds funny, yes, but why is that absurd? Insert anything else in place of the words "government regulation" -- like, oh, "my kid" or "the construction company" -- and it doesn't seem absurd anymore.
Perhaps, but now we're out beyond what Ubuntu offers too.
Other than the fact that running multiple VMs is not going to cost you any more than running one Ubuntu with no VM.
Either I have to jump through hoops when my software doesn't work, or I get something firewall popup spam.
Firewall popup spam would tend to suggest that something is wrong. The XP firewall seems to do something like this, by default, and it is nowhere near as annoying as UAC -- partly because almost no programs actually need to connect to the Internet, or even try, compared to programs which have assumptions about being the admin.
I wouldn't need isolated profiles or VMs then either if I'm just worried about getting 'trashed', if we can assume I can just restore from a backup, then it might as well be a baremetal restore from my backup software live cd, which only takes an hour or so.
Because a snapshotting/copy-on-write VM can switch to the backup in more like a minute, and during that time, you can keep working in the others.
But my point is, why bother just restoring the profile, when restoring another 10GB of OS+Programs is really a drop in the bucket.
That assumes a single profile with some 250 gigs of data.
Its really not unique to unix.
True, although to my knowledge, chroot is. I'm not sure I've seen a Windows app as locked down as, say, Postfix...
The only reason ubuntu is better of than Vista is that it doesn't have the legacy ecosystem of malware and misbehaved applications.
Fair enough, given this context.
I can certainly find other reasons Ubuntu is better -- off the top of my head, there was a Vista bug in which each keystroke typed into the start menu's search bar would spawn a separate thread to run the search. With a good index, it was usually fine, but if you turn off indexing, it's not only going to slowly grind through your hard drive, it's going to start a thread to do that with every keystroke typed. Typing a word into the search bar brings your computer to its knees.
Granted, Ubuntu doesn't have search right away, but Kubuntu does come with Katapult, and it is possible to install Beagle, which is not quite so stupid with threading.
That's an argument in favor of Ubuntu to be sure, but the 'pain' experienced by windows users with Vista is largely a temporary transition issue, as it matures to what we've already got in Ubuntu in terms of security.
You're assuming users actually do switch to Vista.
Regardless, I do think it would have been better to throw all legacy apps under some sort of a VM. When OS X came out, yes, there was Classic, but no one targeted it for new apps -- so I think your argument about it being a "temporary transition issue" would be true of VMs, too. And a full VM only being required for legacy apps is also an annoyance with said apps, probably enough to get their authors to release an updated version -- but I don't think it would be as annoying as UAC.
Anyone care to explain to me why a completely informal, unenforced "Bill of Rights", between Comcast and whatever commercial entities exist in P2P, is any better for consumers than government intervention?
Or answer this: If Comcast really is willing to cooperate, why are they so terrified of government regulation? Why is a legally mandated "Bill of Rights" worse for them than what they are proposing?
The obvious answer is, if it was a law, they couldn't simply violate it.
Next question: Why is Comcast working with BitTorrent, the company? Why do they need to "work with" any P2P corporations, rather than simply dropping their packet shapers and letting P2P protocols work well? Smells to me like Microsoft cutting a deal with Novell -- Microsoft obviously can't cut a deal with Linux itself, as it's a completely distributed, fault-tolerant community, so there's no one CEO to buy -- so they make a deal with Novell, while leaving everyone else out in the cold. Smells to me like Comcast is trying to do the same with P2P -- they can't make a deal with every single filesharer, everywhere, and they won't accept simply falling back to net neutrality, which is what we really want -- so they make a deal with some company which does filesharing, leaving everyone else out in the cold.
That may be true, but really, if I have search-assistant-from-hell dropping popups, and redirecting links, how does it really benefit me if its running in a VM?
In that case, mostly because there are more and better tools for VMs to simplify the process of rolling back to a known-good state, or of having one central known-good state which others may deviate from.
Specifically: You could add one button which would take about a minute to completely nuke the VM, and restore from a known-working version. Depending on how cool your VM software is, that likely includes the resume time.
captured.
If my user profile has been been botted and is participating in ddos and spam attacks at someone else whime, how does it really benefit me if its running in a virtualized environment. Its still doing its job.
Combine that VM with a firewall. Now it's going to ask you for permission, at the very least, to spam.
If my user profile gets ransacked by a virus and all my files and accounts are stolen or trashed...
Can't be stolen if the network is sufficiently locked down. Can't be trashed, because you are keeping backups, right?
Sure the 'system' is still clean, and that's worth something when its time to clean up the mess
Both the host, and any other VMs you're running. The idea is that you aren't running everything in a VM, but rather, that you run things in a VM which need to be run in a VM (for whatever reason), and you compartmentalize.
Take the example of user accounts. It's long been said that Unix was more secure because of limited user accounts, but that's only half the story -- the real reason was that apps would generally behave well if you ran them under another account, and just about all daemons would have their own accounts to run under. Thus, if someone takes over my Postfix server, somehow, they're still not going to get my webserver -- more damage has been done than I would like, yes, but the entire machine isn't destroyed.
Unfortunately, I don't think anyone's really researched how to make sandboxing accessible to the user. Just about any videogame I'm playing needs absolutely no information out of my user profile -- it might benefit from some, like my name, preferred keymap, etc, but there's no reason Portal needs to know about Quakewars, and vice versa. But to put them under separate accounts or VMs is a fair amount of effort right now, so it isn't done very often.
Regardless, even if running only a single VM, there are advantages.
That said, the functionality is there the only question is the price, and that's par for the course on a proprietary system. Even Apple stopped bundling classic long before a lot of people were done needing it.
I wonder if it might be a good idea for them to start releasing source code to versions they no longer intend to support, or no longer wish to provide backwards compatibility for. It's done wonders for Doom.
But what I have noticed is too things.... For example, who can foreget the old nightmare days if configuring soundcards or interupts on PCs[ed: take a deep breath, there should be a comma here] and the difficulty of finding software that worked with your card. Macs all had (somewhat) high end sound cards from very early days[ed: Another comma!] and the driver's for them in the OS distro. So developers could assume they existed.
As a result[comma!] even though I might not actually need some cheerful toon in some piece of software I bought....
nice... this meant my macs had longer service lifetime[comma!] because I was not going and trying to find comaptiblilty extensions and drivers. the old macs had them.
the only place where ala-carte specing...
I don't usually bring it up, but that has to be probably one of the better comments that I've read, in terms of actual content and meaning, with perhaps the worst error/paragraph ratio outside of MySpace! Seriously, a little study would make it a lot easier to read. It's for your own good.
And as a response: I upgrade frequently enough that if I missed something that I actually wanted this time around, I'll get it next time. And all those little things tend to add up to more than I want to spend, just about every time I've looked at a Mac.
That, and they don't have a low-end offering, at all. There's really not much out there to compete with the EEE PC.
1) Microsoft has LONG put out guides on how the software SHOULD work, where it SHOULD store its files and settings.
Which they don't follow themselves.
Sure Vista could virtualize the registry, virtualize the file system, and run legacy software in emulation, and not throw up the uac spam... but all that would result in is the very malware they are trying to block would now be "compatible" with Vista.
...In a virtual machine, where it can't really do much harm. And Windows XP already had a mode whereby every new piece of software attempting an outbound connection would pop up a prompt.
In Linux, if I dug up some old legacy app that expected to be run as root and tried running it, it would just die.
Right, if you just tried running it as a normal user, with nothing changed. Linux doesn't have anything built-in to tell it that this app is old and needs to be treated specially.
It is, however, trivial to fool an app into thinking it has root -- Debian provides a "fakeroot" utility to do just that. Combined with a chroot, it may even be possible to run this app without attempting to patch it in strange ways.
2) They do give away Virtual PC 2007, so if you really want XP emulation, you can have it, in a nice nearly completely isolated sandbox.
And emulate an entire computer? Yuck. How's virtual PC at 3D acceleration, by the way?
No, I meant something closer to Wine, but sandboxed -- in other words, something very much like Classic. There's no reason for it to need an entire OS in there, just enough for old apps to work.
Also: They give Virtual PC 2007 away, but they don't give XP away. This means you can't buy a Vista upgrade, you have to by the whole thing -- or, if you buy a new computer, you have to buy XP separately.
And they want it to launch when they plug their camera in, and do everything in one place.
At the very least, cameras are now detected and treated as cameras. The process isn't as seamless, but it's pretty close -- copy them with digiKam, which prompts you on first run to tell it where to save photos -- so if you discover Krita later, you know where they are. Analogous to downloading. Most users don't even need that -- rotate, flip, turn to black and white, resize, and some batch modifications are built right in to digiKam. And that's all without having a digital camera handy, and just reading the disabled menu entries -- looking at the docs, it looks as though there's more image editing there.
I didn't know any of this -- I'm the kind of person who writes a script with a commandline camera utility, when I actually have a camera -- but it's far from a bad experience, especially considering the camera is, in fact, plug'n'play (with this software already there), whereas on Windows, there's a chance they'd have to install drivers.
Like all other Dell support it covers what you bought from Dell.
Probably. I seem to remember that it's provided by Canonical, though.
Interestingly, when using slightly different tools, but in the same process and for the same need, it usually is covered by the patent.
That's why I see them as holding back progress, especially when they last for so long.
Fifteen years ago, Windows 95 had just come out, and people were starting to get used to the idea of the Internet. Imagine someone had managed to patent the idea of setting up a universal directory of websites, searchable by keywords, and indexed for efficiency. Where would Google be today?
I find that it's precisely the things which aren't patented, or the areas in which patents are ignored, that the more interesting things happen. I have a hard time imagining a patent system where this is not the case.
As a result, when any layman performs the overtness test, they try and figure out if the need is unique first, instead of what you're doing: looking at the technologies and implementations for cleverness.
It's more that I feel that both the technology and the implementation is obvious.
Specifically: Given the need to fastforward a finite amount of time, an implementation of creating a button for that, and seeking through an index in the stored video, is kind of obvious. Given a desire to skip commercials, and the reality of what people were already doing with VCRs, the need to fastforward a finite amount of time seems obvious.
It could be that hindsight is 20/20 there, or that I really need to actually read that specific patent.
Mathematics is a troubled area. It can't be both copyrightable and patentable; it cannot be both painting and invention.
The same would be true of software, I would think.
Actually, I don't entirely agree there. A painting may not be functional, but it is possible for something to be both functional and a beautiful work of art. The trick is, in the real world, it's much easier to figure out what is form and what is function -- in software, not so much.
Say you're editing an image that involves looking at a spreadsheet program side-by-side with the image. (For instance, you're pretty-ify-ing a set of charts.) A virtual desktop can't help you in this situation...
Well, assuming the spreadsheet program is a typical one, and sticks to one (possibly tabbed) window, there's actually a very simple answer to this: In KDE, I have middle-clicking on the taskbar set to send a window to the bottom of the stack.
So, not only do I have more freedom in arranging my stuff around the spreadsheet -- for example, I can put the image on one monitor, and tools+spreadsheet on the other monitor -- but I also have that alt-tab-ish functionality back, where I can simply send the spreadsheet to the bottom, which, since it's the only other thing on the Gimp virtual desktop, means all the Gimp windows go to the top.
GIMP does it a different way for no good reason
You said that "Virtual desktops doesn't help in a lot of situations" -- which implies that there are some in which it does help. Your example of one in which virtual desktops don't help is actually a situation where they do. I'd call that a very good reason.
GIMP developers are utterly clueless about the whole affair.
I think it's more that they know, and don't care. Their way is better, in many situations. And if you really don't like it, there are ways to force Gimp back to a single window.
If your shell script is not POSIX/bin/sh, don't mark it as a POSIX/bin/sh script. Is that difficult?
If it's not a shell script, and not mine, then yes. I have to crawl through and find every "system" call that looks suspiciously bash-like... Or I have to temporarily relink/bin/sh, or run it in a chroot.
FWIW The trick to use dash as/bin/sh was well known by loads of Debian users back in the day when it was still called "ash".
Just pointing out that Ubuntu forcing this on users pretty much ended the random scripts people would have that were lazy about it. Except for things like what I just described -- Amazon's EC2 image-building tools are written in Ruby, with the assumption of/bin/sh as bash.
The use of it in N770 lead to loads of scripts being made clean of bashisms. But now with the popularity of Ubuntu, many, many people are forcing "#!/bin/bash" down everyones throats even in cases when it is not necessary.
Kind of is -- or at least, I would rather have a script explicitly request Bash and work, than use one or two bashisms it doesn't need and fail.
In any case, currently "grep bin/bash/etc/init.d/* | wc -l" gives me a count of 5. Why? I mean why on the scripts I use to boot? Five stupid unnecessary reasons for my computer to boot slower, and to force me to have bash installed no matter what.
You're always free to look in there and see if you can do it better.
Personally, I use bash anyway, as an interactive shell. And if I ever run into a Bashism that I absolutely must have in a script, then I absolutely will make it require Bash -- just as if I need something more complicated than a shell script, I'll make it require Ruby or Python.
Additional cons:
- Most often curved. Some claim to have a flat screen, but don't. This distorts the picture, sometimes ridiculously slow.
- Harder to calibrate. H-Position, V-Position, Degauss, etc. On an LCD, hit "auto-adjust" and done.
- Bigger than an LCD. Takes up more desk space.
- There is no matte. They're all glossy -- glass, actually -- and they can glare something fierce.
- No DVI or HDMI input, as far as I can tell.
- Requires a warm-up time. Alternatively, the picture starts out bigger, and slowly shrinks, requiring you to re-adjust it to fit the screen.
- Flicker.
- Blurry. Perhaps they can be adjusted to be better, but a decent LCD is always sharp. And if it's not, "auto-adjust" and done.
That about sums up why I don't really want to go back to CRT. But I do wish I could find some decent LCDs which don't take shortcuts.
I fully admit to being in the minority who enjoys it. But I do think it's useful to know, because then you get to do shell scripts. GUIs are not really scriptable.
You think the internet boom would of been such a big pop if people weren't so backed by cheap hosting provided by Linux-power?
There are many things that contributed to the Internet being a big deal. Linux is only part of it. (And if it wasn't Linux, it would have been BSD...)
But I would say that it would have been about as big, anyway. There was an insane amount of money coming from venture cap firms, etc -- everyone had a dotcom, and they all had Underpants Gnomes business models. Cheap Linux may have accelerated it, but requiring expensive Windows (or Solaris) really just means more money.
But that's what these are, single-mission set-ups! There's a service to be provided, a task to be performed, and it can do this great.
Yes. So can most OSes now -- we're past the memory leak, bluescreen every five hours BS.
Not general purpose, do everything, run a bunch of crap that this guy who knows nothing about what a command prompt is...
Which has nothing to do with running more than one task, and doing it well. I think you've got multitasking and user-friendliness confused, and I have no idea how you got them confused.
Trivial example: Xen virtualization. Amazon EC2 does this now, as does Slicehost. It means that a single physical Linux machine may be running some 30 virtual machines, each of which, if its admin is particularly unimaginative, is only doing one thing.
And for that matter, Amazon EC2 is establishing itself as a general-purpose, do-anything, CPU-on-demand solution.
and barely feels confident about clicking all those 'Next' buttons when installing an app.
That should be a point for Linux, not against it. True, some apps are going to require manual installs that are far worse than anything Windows has...
And others are going to be packages, which can be installed in a few clicks from the package manager, or a very simple command on the command-line. No "next" buttons in sight. It may be more complex, the first time -- but so was a next-next-next wizard, until users figured out they could ignore everything and just keep hitting next until it was done. And once the user knows it, they can confidently install or uninstall anything in the repository.
Most end-users do not have a single mission or purpose.
Again: Neither do all Linux servers.
Windows was coded from the ground-up for multitasking business work...
Windows wasn't even coded from the ground up; large chunks of it were stolen from Apple, originally.
And it's also not entirely accurate -- it took long enough for businesspeople to quite grasp the concept of multitasking on Windows.
I didn't mean to convey GUI-code is inside the Windows kernel (I don't know). Simply that Windows integrates the modular NT structure very seamlessly, resulting in the product called Windows.
Funny you should say that. I find that Kubuntu/KDE integrates a ton of projects, made by completely different people, fairly seamlessly, into one product.
It's also not why people use Windows. Ironically, desktop users continue to use Windows largely because of third-party software, which is not necessarily integrated very well. Linux may not be as "tightly integrated", whatever the hell that means (I think you mean consistent, navigateable GUIs), but the things it does integrate, it integrates pretty much as if they were third-parties.
The main irritation I have with EC2 is that it's too low-level, but it does mean it can run just about anything, including App Engines.
App Engines will not, however, be able to run EC2. (Kind of obvious, if you know anything about either of them.)
However, I think you lose the main benefit of using App Engines if you put them on EC2 -- that being that Google gets to worry about scaling. With EC2, you have to do everything yourself, including detecting load and deciding whether or not to fire up another instance. With App Engines, you just upload your app and watch it go, unless I'm misunderstanding something. Put App Engines on EC2, and you suddenly have to build an infrastructure to support it.
So it's nice to know your app is portable, at least, but I don't think anyone's seriously suggesting this, other than as a way to keep Google on their toes -- if Google really does start to be evil, this is a nice way to port away from them.
Either you have no idea what you're talking about, or you're not communicating very well...
There's no standardization for interaction. You have different shells, different window managers, different distros.
Interaction between what?
There are many standards. POSIX provides standards for shells -- a shell must support a certain set of features to be POSIX-compliant, and a POSIX-compliant shell script can run with #!/bin/sh on any remotely POSIX-compliant system.
Without it much of the web wouldn't exist as it now does, and I see Linux as having been crucial in making it better, as well as sadly making it worse.
I don't really get how Linux is responsible for either of these things. All I see is cheap, commodity webservers, which made it easier for a startup to get somewhere. Remember, even Google was a startup not very long ago.
Given the powerfulness of the kernel to tackle complicated tasks, but usually is best fitted for single-mission services.
WTF?
What do you mean by a "complicated task"? And what do you mean by "single-mission services"?
Compared with a mature GUI model
X was released in 1984. Nineteen eighty-fucking-four. If you are less than 25 years old, Linux's GUI system is more mature than you.
with a tight integration with the kernel
Why is this a good thing?
Performance? Depending on the drivers, Linux has frequently beat Windows at raw 3D performance for awhile now.
Other than that, what is the point? Why would I want all that GUI crap in my kernel? Seems like even Microsoft has started to figure out that this is a Bad Idea, and newer versions of Windows are splitting more and more drivers away from the kernel -- the next version of Windows Server should be able to boot without a GUI.
a highly sophisticated, and highly extensible threading model
What does that even mean?
Linux has threads. Linux also has lightweight fork()s. Linux can spawn new, whole processes faster than Windows or OS X -- or any other OS I know of.
as well as other modular subsystems
I'm sorry, "tight integration with the kernel" is the opposite of modular. Tight integration is the opposite of modularity, full stop.
I wouldn't even dare hosting a critical web service on Windows, regardless of version.
Many people do, now. Windows is now a viable server platform. There are still many good reasons for preferring Linux, or any Unix, but to say you wouldn't "dare" suggests that you see some security risk that just isn't relevant anymore -- Windows can be locked down as tightly as any Linux, and the more dangerous vulnerabilities are usually in the application, now.
Let me ask you this: When was the last time Microsoft.com was 0wned?
We believe standards debate should always be carried out with respect for all parties, even when they strongly disagree.
Respect is earned.
This is not a matter of "strongly disagree" -- in about ten minutes, I was able to explain to my mother why OOXML is not particularly open, and not feasible to implement by third-parties -- and thus should not be a standard. Yes, my mother understands this, but the ISO doesn't.
Even disregarding that -- even assuming that's what they meant by "strongly disagree" -- Microsoft has thoroughly gamed the process, something which, again, is easy to see.
All of which means that personal attacks are entirely justified -- but it looks as though there weren't any:
The plenary session was marked by protests outside, largely carried out by delegates from a nearby open-source conference. The protesters were calling for OOXML to be withdrawn from ISO standardisation something that could theoretically happen if a national standards body were to protest against its own vote within the next month or two.
That's right -- a protest.
In what way is a protest a "personal attack"? It is an attack against your organization, or against your working group -- it is not an attack against you, personally, and it is not an attack against anything about you other than that you passed a standard which, to put it kindly, is obviously incomplete and broken.
Speaking of which -- you brought this on yourselves. You passed a standard which is obviously incomplete and broken -- or, in mrchaotica's words, it is redundant, conflicting, and poorly-designed. You absolutely deserve every single flame, troll, protest, and dirty look that you get. I, and others like me, are going to do our best to make it at least as unpleasant for you to pass OOXML, as Microsoft would endeavor to make it unpleasant for you to not pass OOXML.
Last I checked, apple-tab on a mac would cycle through apps, not windows. It would thus raise all Gimp windows when you find the Gimp, as a single item. Maybe I had to tweak something to turn that behavior on, but that's pretty much how I work when I'm on OS X -- on Gimp or anything else.
The menus are just... oddd. To a new user, the app is useless because once you open something, you get a window with no menus. After much frustration, the user monkey-clicks the mouse and realizes the menus are on the right-click instead of at the top of the window.
Apparently, you haven't used it in 2 years. The image windows now have their own menu bar, which pretty much replaces right-click.
I always found using multiple windows was a good idea -- it lets the window manager actually manage the windows. If it's annoying, in the ways you describe, maybe that says something about your window manager?
Now the reason I think you're wrong is because last time I posted a rant like this, two Mac users pointed out that Photoshop on the Mac does not work like this. Apperently, clicking on any image open in photoshop also brings the tools into focus as well.
At least on OS X, that is how all programs work. Or at least, it is possible to click on an application to raise all of its windows, and command+tab (like alt+tab, but better) will actually raise all of those windows. Windows are actually naturally grouped by application -- I had a keystroke to cycle through open Terminals, and that actually worked really well, because Terminal is actually its own application.
Gimp was developed on Linux, where we've had a few sane windowing ideas that Windows has yet to pick up on, and OS X is only slowly starting to steal. Simple example: Virtual desktops. Put gimp on its own workspace, and you are literally one keystroke away from moving back to that image.
And then there are dual-monitor systems. This is where Photoshop really starts to be annoying, unless there is some way I don't know of to detach the tools (probably is) -- it's possible to put the image itself, completely maximized, on one monitor, and all of the tools on another monitor.
Clicking an image in Gimp should bring the image up along with a full set of editing tools.
Most open source programs try to assume less about their user -- what if you didn't want that full set of editing tools to come up? What if you just want to look at the image, on as much screen area as possible, before you start editing? Why should it be the job of the individual application to work around crappy window managers?
All that said, there's always GimpShop -- haven't tried it myself, but it claims to make Gimp look photoshop-like.
Debian doesn't seem to mind Ubuntu, for the most part, except when Ubuntu users come to Debian forums asking for help.
But a simple example: Debian, for the longest time, had/bin/sh linked to/bin/bash, although they did have a rule that any script requiring/bin/sh should only use POSIX syntax, and not bash-isms. Sometime in 2006, I think, Ubuntu switched/bin/sh to/bin/dash. Dash is much faster than Bash -- so much so that this switch is the main reason that version of Ubuntu booted so much faster than previous versions (it was also when Upstart was first integrated, though Upstart is barely used)...
And since then, certainly, fixes to various packages' scripts which claim #!/bin/sh, but really want bash, have been sent back to Debian. (Either POSIX-ify them, or make them explicitly ask for bash.) But as far as I know, no major distributions outside Ubuntu actually have/bin/sh linked to dash -- some of Amazon's EC2 tools, designed to work on Fedora, need to be patched before they can work on Ubuntu, for that very reason.
Aside from what happens when I actually have a movie to watch, there's the fact that I often can find a few columns to fill that space.
Current example: Kate, a text editor, with a filesystem browser on the left, and a list of open documents on the right, with the actual editing space between them. Maximize it, and I don't feel like any space is wasted. Certainly not as good as if I had a full 1920x1200, or 1920x1080, but I can usually find a way to fill any decent screen.
Slashdot Mirror
Last I checked, Mandriva was proprietary. It may have a free version, but it depends on proprietary software. Right?
Ubuntu, by default, comes with entirely free software. The proprietary stuff is easy to install, but it's nice to know that most of the core stuff is free, and will always be free.
Is there anything Mandriva has over Ubuntu?
Dell sells fairly stock, high-end desktops with Ubuntu preloaded. When buying a new computer, at least, you can let Dell go through all the pain of installing and getting it to work with your hardware -- and when it doesn't, you get a support contract.
Which is, by the way, exactly what you get from Dell when you buy Windows.
Linux does support software RAID, and many RAID controllers. But realize that by using hardware RAID, you are well beyond "the masses", and should be prepared to look this stuff up yourself.
Also, the server and alternate install CDs are much more likely to work with RAID and similar things, just as the desktop install CDs are much more likely to setup X correctly. If you know what you're doing, though, you can simply install from the server CD and pull in the missing components from the desktop.
What brand is it?
This is a place where support is spotty lately. That's not an excuse, and not a way of saying it's "ready", but if you are interested, it should be pretty simple to find a list of supported hardware, including which wireless cards work the best. Anything Intel is probably going to work out of the box, with open drivers.
And for the slightly less wealthy masses, it's called the EEE PC.
Last time I installed Windows, it always warned when formatting. That is: It won't delete a partition without requiring a series of steps which are clearly designed to avoid the user just hitting next-next-next, yes-yes-yes, or reflexively doing whatever they think they need to do to procede.
The same is true of formatting -- the XP installer, at least, clearly warns that you won't be able to recover.
Now, of course, Ubuntu makes it incredibly easy to shrink your existing Windows install, whereas Windows provides no tools for dealing with other OSes at all (or that I know of). But it does warn you every step of the way, in typical Windows fashion, that you had better be damned-fucking-sure before you nuke any data.
As a power user, I would love for Linux to be mainstream. The more mainstream it gets, the more likely my video drivers are to work, and the more likely I am to have some decent games to play.
As a server administrator, I would love it if all of our developers ran Linux on their desktops. It's still possible to run into surprises deploying from Windows on their workstations (read: laptops) to Linux on the server.
Here's the cool part: It's not up to you.
The thing is, Linux -- or, more generally, all open source software -- is for everything and everyone. If there's anyone who can't use it, or anything it can't yet do, that's just another problem to be fixed by anyone who has the time.
And no one can stop it. You can't make it into your 31337 high-school h4x0r club anymore. It's much bigger than that, now.
Whenever I hear this, I always ask: Are you seriously suggesting that there be more than one company in a given area running physical cables to every house? Or are you suggesting more government regulation to force them to share the cables they've got?
It sounds funny, yes, but why is that absurd? Insert anything else in place of the words "government regulation" -- like, oh, "my kid" or "the construction company" -- and it doesn't seem absurd anymore.
Other than the fact that running multiple VMs is not going to cost you any more than running one Ubuntu with no VM.
Firewall popup spam would tend to suggest that something is wrong. The XP firewall seems to do something like this, by default, and it is nowhere near as annoying as UAC -- partly because almost no programs actually need to connect to the Internet, or even try, compared to programs which have assumptions about being the admin.
Because a snapshotting/copy-on-write VM can switch to the backup in more like a minute, and during that time, you can keep working in the others.
That assumes a single profile with some 250 gigs of data.
True, although to my knowledge, chroot is. I'm not sure I've seen a Windows app as locked down as, say, Postfix...
Fair enough, given this context.
I can certainly find other reasons Ubuntu is better -- off the top of my head, there was a Vista bug in which each keystroke typed into the start menu's search bar would spawn a separate thread to run the search. With a good index, it was usually fine, but if you turn off indexing, it's not only going to slowly grind through your hard drive, it's going to start a thread to do that with every keystroke typed. Typing a word into the search bar brings your computer to its knees.
Granted, Ubuntu doesn't have search right away, but Kubuntu does come with Katapult, and it is possible to install Beagle, which is not quite so stupid with threading.
You're assuming users actually do switch to Vista.
Regardless, I do think it would have been better to throw all legacy apps under some sort of a VM. When OS X came out, yes, there was Classic, but no one targeted it for new apps -- so I think your argument about it being a "temporary transition issue" would be true of VMs, too. And a full VM only being required for legacy apps is also an annoyance with said apps, probably enough to get their authors to release an updated version -- but I don't think it would be as annoying as UAC.
Anyone care to explain to me why a completely informal, unenforced "Bill of Rights", between Comcast and whatever commercial entities exist in P2P, is any better for consumers than government intervention?
Or answer this: If Comcast really is willing to cooperate, why are they so terrified of government regulation? Why is a legally mandated "Bill of Rights" worse for them than what they are proposing?
The obvious answer is, if it was a law, they couldn't simply violate it.
Next question: Why is Comcast working with BitTorrent, the company? Why do they need to "work with" any P2P corporations, rather than simply dropping their packet shapers and letting P2P protocols work well? Smells to me like Microsoft cutting a deal with Novell -- Microsoft obviously can't cut a deal with Linux itself, as it's a completely distributed, fault-tolerant community, so there's no one CEO to buy -- so they make a deal with Novell, while leaving everyone else out in the cold. Smells to me like Comcast is trying to do the same with P2P -- they can't make a deal with every single filesharer, everywhere, and they won't accept simply falling back to net neutrality, which is what we really want -- so they make a deal with some company which does filesharing, leaving everyone else out in the cold.
Gotta love the smell of bullshit in the morning.
In that case, mostly because there are more and better tools for VMs to simplify the process of rolling back to a known-good state, or of having one central known-good state which others may deviate from.
Specifically: You could add one button which would take about a minute to completely nuke the VM, and restore from a known-working version. Depending on how cool your VM software is, that likely includes the resume time.
captured.Combine that VM with a firewall. Now it's going to ask you for permission, at the very least, to spam.
Can't be stolen if the network is sufficiently locked down. Can't be trashed, because you are keeping backups, right?
Both the host, and any other VMs you're running. The idea is that you aren't running everything in a VM, but rather, that you run things in a VM which need to be run in a VM (for whatever reason), and you compartmentalize.
Take the example of user accounts. It's long been said that Unix was more secure because of limited user accounts, but that's only half the story -- the real reason was that apps would generally behave well if you ran them under another account, and just about all daemons would have their own accounts to run under. Thus, if someone takes over my Postfix server, somehow, they're still not going to get my webserver -- more damage has been done than I would like, yes, but the entire machine isn't destroyed.
Unfortunately, I don't think anyone's really researched how to make sandboxing accessible to the user. Just about any videogame I'm playing needs absolutely no information out of my user profile -- it might benefit from some, like my name, preferred keymap, etc, but there's no reason Portal needs to know about Quakewars, and vice versa. But to put them under separate accounts or VMs is a fair amount of effort right now, so it isn't done very often.
Regardless, even if running only a single VM, there are advantages.
I wonder if it might be a good idea for them to start releasing source code to versions they no longer intend to support, or no longer wish to provide backwards compatibility for. It's done wonders for Doom.
I don't usually bring it up, but that has to be probably one of the better comments that I've read, in terms of actual content and meaning, with perhaps the worst error/paragraph ratio outside of MySpace! Seriously, a little study would make it a lot easier to read. It's for your own good.
And as a response: I upgrade frequently enough that if I missed something that I actually wanted this time around, I'll get it next time. And all those little things tend to add up to more than I want to spend, just about every time I've looked at a Mac.
That, and they don't have a low-end offering, at all. There's really not much out there to compete with the EEE PC.
Which they don't follow themselves.
...In a virtual machine, where it can't really do much harm. And Windows XP already had a mode whereby every new piece of software attempting an outbound connection would pop up a prompt.
Right, if you just tried running it as a normal user, with nothing changed. Linux doesn't have anything built-in to tell it that this app is old and needs to be treated specially.
It is, however, trivial to fool an app into thinking it has root -- Debian provides a "fakeroot" utility to do just that. Combined with a chroot, it may even be possible to run this app without attempting to patch it in strange ways.
And emulate an entire computer? Yuck. How's virtual PC at 3D acceleration, by the way?
No, I meant something closer to Wine, but sandboxed -- in other words, something very much like Classic. There's no reason for it to need an entire OS in there, just enough for old apps to work.
Also: They give Virtual PC 2007 away, but they don't give XP away. This means you can't buy a Vista upgrade, you have to by the whole thing -- or, if you buy a new computer, you have to buy XP separately.
At the very least, cameras are now detected and treated as cameras. The process isn't as seamless, but it's pretty close -- copy them with digiKam, which prompts you on first run to tell it where to save photos -- so if you discover Krita later, you know where they are. Analogous to downloading. Most users don't even need that -- rotate, flip, turn to black and white, resize, and some batch modifications are built right in to digiKam. And that's all without having a digital camera handy, and just reading the disabled menu entries -- looking at the docs, it looks as though there's more image editing there.
I didn't know any of this -- I'm the kind of person who writes a script with a commandline camera utility, when I actually have a camera -- but it's far from a bad experience, especially considering the camera is, in fact, plug'n'play (with this software already there), whereas on Windows, there's a chance they'd have to install drivers.
Probably. I seem to remember that it's provided by Canonical, though.
That's why I see them as holding back progress, especially when they last for so long.
Fifteen years ago, Windows 95 had just come out, and people were starting to get used to the idea of the Internet. Imagine someone had managed to patent the idea of setting up a universal directory of websites, searchable by keywords, and indexed for efficiency. Where would Google be today?
I find that it's precisely the things which aren't patented, or the areas in which patents are ignored, that the more interesting things happen. I have a hard time imagining a patent system where this is not the case.
It's more that I feel that both the technology and the implementation is obvious.
Specifically: Given the need to fastforward a finite amount of time, an implementation of creating a button for that, and seeking through an index in the stored video, is kind of obvious. Given a desire to skip commercials, and the reality of what people were already doing with VCRs, the need to fastforward a finite amount of time seems obvious.
It could be that hindsight is 20/20 there, or that I really need to actually read that specific patent.
The same would be true of software, I would think.
Actually, I don't entirely agree there. A painting may not be functional, but it is possible for something to be both functional and a beautiful work of art. The trick is, in the real world, it's much easier to figure out what is form and what is function -- in software, not so much.
And higher price.
Yes.
That's the idea. That, and maybe getting the other companies to lean on this one.
Well, assuming the spreadsheet program is a typical one, and sticks to one (possibly tabbed) window, there's actually a very simple answer to this: In KDE, I have middle-clicking on the taskbar set to send a window to the bottom of the stack.
So, not only do I have more freedom in arranging my stuff around the spreadsheet -- for example, I can put the image on one monitor, and tools+spreadsheet on the other monitor -- but I also have that alt-tab-ish functionality back, where I can simply send the spreadsheet to the bottom, which, since it's the only other thing on the Gimp virtual desktop, means all the Gimp windows go to the top.
You said that "Virtual desktops doesn't help in a lot of situations" -- which implies that there are some in which it does help. Your example of one in which virtual desktops don't help is actually a situation where they do. I'd call that a very good reason.
I think it's more that they know, and don't care. Their way is better, in many situations. And if you really don't like it, there are ways to force Gimp back to a single window.
If it's not a shell script, and not mine, then yes. I have to crawl through and find every "system" call that looks suspiciously bash-like... Or I have to temporarily relink /bin/sh, or run it in a chroot.
Just pointing out that Ubuntu forcing this on users pretty much ended the random scripts people would have that were lazy about it. Except for things like what I just described -- Amazon's EC2 image-building tools are written in Ruby, with the assumption of /bin/sh as bash.
Kind of is -- or at least, I would rather have a script explicitly request Bash and work, than use one or two bashisms it doesn't need and fail.
You're always free to look in there and see if you can do it better.
Personally, I use bash anyway, as an interactive shell. And if I ever run into a Bashism that I absolutely must have in a script, then I absolutely will make it require Bash -- just as if I need something more complicated than a shell script, I'll make it require Ruby or Python.
Additional cons:
- Most often curved. Some claim to have a flat screen, but don't. This distorts the picture, sometimes ridiculously slow.
- Harder to calibrate. H-Position, V-Position, Degauss, etc. On an LCD, hit "auto-adjust" and done.
- Bigger than an LCD. Takes up more desk space.
- There is no matte. They're all glossy -- glass, actually -- and they can glare something fierce.
- No DVI or HDMI input, as far as I can tell.
- Requires a warm-up time. Alternatively, the picture starts out bigger, and slowly shrinks, requiring you to re-adjust it to fit the screen.
- Flicker.
- Blurry. Perhaps they can be adjusted to be better, but a decent LCD is always sharp. And if it's not, "auto-adjust" and done.
That about sums up why I don't really want to go back to CRT. But I do wish I could find some decent LCDs which don't take shortcuts.
Which is where X comes in.
I fully admit to being in the minority who enjoys it. But I do think it's useful to know, because then you get to do shell scripts. GUIs are not really scriptable.
There are many things that contributed to the Internet being a big deal. Linux is only part of it. (And if it wasn't Linux, it would have been BSD...)
But I would say that it would have been about as big, anyway. There was an insane amount of money coming from venture cap firms, etc -- everyone had a dotcom, and they all had Underpants Gnomes business models. Cheap Linux may have accelerated it, but requiring expensive Windows (or Solaris) really just means more money.
Yes. So can most OSes now -- we're past the memory leak, bluescreen every five hours BS.
Which has nothing to do with running more than one task, and doing it well. I think you've got multitasking and user-friendliness confused, and I have no idea how you got them confused.
Trivial example: Xen virtualization. Amazon EC2 does this now, as does Slicehost. It means that a single physical Linux machine may be running some 30 virtual machines, each of which, if its admin is particularly unimaginative, is only doing one thing.
And for that matter, Amazon EC2 is establishing itself as a general-purpose, do-anything, CPU-on-demand solution.
That should be a point for Linux, not against it. True, some apps are going to require manual installs that are far worse than anything Windows has...
And others are going to be packages, which can be installed in a few clicks from the package manager, or a very simple command on the command-line. No "next" buttons in sight. It may be more complex, the first time -- but so was a next-next-next wizard, until users figured out they could ignore everything and just keep hitting next until it was done. And once the user knows it, they can confidently install or uninstall anything in the repository.
Again: Neither do all Linux servers.
Windows wasn't even coded from the ground up; large chunks of it were stolen from Apple, originally.
And it's also not entirely accurate -- it took long enough for businesspeople to quite grasp the concept of multitasking on Windows.
Funny you should say that. I find that Kubuntu/KDE integrates a ton of projects, made by completely different people, fairly seamlessly, into one product.
It's also not why people use Windows. Ironically, desktop users continue to use Windows largely because of third-party software, which is not necessarily integrated very well. Linux may not be as "tightly integrated", whatever the hell that means (I think you mean consistent, navigateable GUIs), but the things it does integrate, it integrates pretty much as if they were third-parties.
Trivial example: Anyone can create a Debian
The main irritation I have with EC2 is that it's too low-level, but it does mean it can run just about anything, including App Engines.
App Engines will not, however, be able to run EC2. (Kind of obvious, if you know anything about either of them.)
However, I think you lose the main benefit of using App Engines if you put them on EC2 -- that being that Google gets to worry about scaling. With EC2, you have to do everything yourself, including detecting load and deciding whether or not to fire up another instance. With App Engines, you just upload your app and watch it go, unless I'm misunderstanding something. Put App Engines on EC2, and you suddenly have to build an infrastructure to support it.
So it's nice to know your app is portable, at least, but I don't think anyone's seriously suggesting this, other than as a way to keep Google on their toes -- if Google really does start to be evil, this is a nice way to port away from them.
Either you have no idea what you're talking about, or you're not communicating very well...
Interaction between what?There are many standards. POSIX provides standards for shells -- a shell must support a certain set of features to be POSIX-compliant, and a POSIX-compliant shell script can run with #!/bin/sh on any remotely POSIX-compliant system.
I don't really get how Linux is responsible for either of these things. All I see is cheap, commodity webservers, which made it easier for a startup to get somewhere. Remember, even Google was a startup not very long ago.
WTF?
What do you mean by a "complicated task"? And what do you mean by "single-mission services"?
X was released in 1984. Nineteen eighty-fucking-four. If you are less than 25 years old, Linux's GUI system is more mature than you.
Why is this a good thing?
Performance? Depending on the drivers, Linux has frequently beat Windows at raw 3D performance for awhile now.
Other than that, what is the point? Why would I want all that GUI crap in my kernel? Seems like even Microsoft has started to figure out that this is a Bad Idea, and newer versions of Windows are splitting more and more drivers away from the kernel -- the next version of Windows Server should be able to boot without a GUI.
What does that even mean?
Linux has threads. Linux also has lightweight fork()s. Linux can spawn new, whole processes faster than Windows or OS X -- or any other OS I know of.
I'm sorry, "tight integration with the kernel" is the opposite of modular. Tight integration is the opposite of modularity, full stop.
Many people do, now. Windows is now a viable server platform. There are still many good reasons for preferring Linux, or any Unix, but to say you wouldn't "dare" suggests that you see some security risk that just isn't relevant anymore -- Windows can be locked down as tightly as any Linux, and the more dangerous vulnerabilities are usually in the application, now.
Let me ask you this: When was the last time Microsoft.com was 0wned?
Respect is earned.
This is not a matter of "strongly disagree" -- in about ten minutes, I was able to explain to my mother why OOXML is not particularly open, and not feasible to implement by third-parties -- and thus should not be a standard. Yes, my mother understands this, but the ISO doesn't.
Even disregarding that -- even assuming that's what they meant by "strongly disagree" -- Microsoft has thoroughly gamed the process, something which, again, is easy to see.
All of which means that personal attacks are entirely justified -- but it looks as though there weren't any:
That's right -- a protest.
In what way is a protest a "personal attack"? It is an attack against your organization, or against your working group -- it is not an attack against you, personally, and it is not an attack against anything about you other than that you passed a standard which, to put it kindly, is obviously incomplete and broken.
Speaking of which -- you brought this on yourselves. You passed a standard which is obviously incomplete and broken -- or, in mrchaotica's words, it is redundant, conflicting, and poorly-designed. You absolutely deserve every single flame, troll, protest, and dirty look that you get. I, and others like me, are going to do our best to make it at least as unpleasant for you to pass OOXML, as Microsoft would endeavor to make it unpleasant for you to not pass OOXML.
This is true.
See, what I heard was, the corn did not directly kill the butterfly, but that the milkweed was dusted with the corn pollen.
The problem is, a quick Google search turns up very little except Monsanto's own press release about the issue.
Or organic farming.
Apparently, you haven't used it in 2 years. The image windows now have their own menu bar, which pretty much replaces right-click.
I always found using multiple windows was a good idea -- it lets the window manager actually manage the windows. If it's annoying, in the ways you describe, maybe that says something about your window manager?
At least on OS X, that is how all programs work. Or at least, it is possible to click on an application to raise all of its windows, and command+tab (like alt+tab, but better) will actually raise all of those windows. Windows are actually naturally grouped by application -- I had a keystroke to cycle through open Terminals, and that actually worked really well, because Terminal is actually its own application.
Gimp was developed on Linux, where we've had a few sane windowing ideas that Windows has yet to pick up on, and OS X is only slowly starting to steal. Simple example: Virtual desktops. Put gimp on its own workspace, and you are literally one keystroke away from moving back to that image.
And then there are dual-monitor systems. This is where Photoshop really starts to be annoying, unless there is some way I don't know of to detach the tools (probably is) -- it's possible to put the image itself, completely maximized, on one monitor, and all of the tools on another monitor.
Most open source programs try to assume less about their user -- what if you didn't want that full set of editing tools to come up? What if you just want to look at the image, on as much screen area as possible, before you start editing? Why should it be the job of the individual application to work around crappy window managers?
All that said, there's always GimpShop -- haven't tried it myself, but it claims to make Gimp look photoshop-like.
Debian doesn't seem to mind Ubuntu, for the most part, except when Ubuntu users come to Debian forums asking for help.
/bin/sh linked to /bin/bash, although they did have a rule that any script requiring /bin/sh should only use POSIX syntax, and not bash-isms. Sometime in 2006, I think, Ubuntu switched /bin/sh to /bin/dash. Dash is much faster than Bash -- so much so that this switch is the main reason that version of Ubuntu booted so much faster than previous versions (it was also when Upstart was first integrated, though Upstart is barely used)...
/bin/sh linked to dash -- some of Amazon's EC2 tools, designed to work on Fedora, need to be patched before they can work on Ubuntu, for that very reason.
But a simple example: Debian, for the longest time, had
And since then, certainly, fixes to various packages' scripts which claim #!/bin/sh, but really want bash, have been sent back to Debian. (Either POSIX-ify them, or make them explicitly ask for bash.) But as far as I know, no major distributions outside Ubuntu actually have
Aside from what happens when I actually have a movie to watch, there's the fact that I often can find a few columns to fill that space.
Current example: Kate, a text editor, with a filesystem browser on the left, and a list of open documents on the right, with the actual editing space between them. Maximize it, and I don't feel like any space is wasted. Certainly not as good as if I had a full 1920x1200, or 1920x1080, but I can usually find a way to fill any decent screen.