Linux Desktop Without X11

A writes "Rocklyte systems have announced the first version of their Athene Operating System. It is a desktop and embedded operating system built on the Linux kernel, but without the "aging X11". Instead, it uses the SciTech SNAP graphics system with which it is possible to completely re-theme the desktop to look like the famous AmigaOS GUI or another famous UI. For backwards compatibility, an X11 server is also available in the system. The system can run completely off the CD, without needing to be installed on the harddrive."

Comments

Themes schemes

[vasqzr]

Instead, it uses the SciTech SNAP graphics system with which it is possible to completely re-theme the desktop to look like the famous AmigaOS GUI or another famous UI.

Right. Because themes are the most important thing, ever. This isn't an media player, it's a GUI.

Re:Themes schemes

[NanoGator]

"Right. Because themes are the most important thing, ever. This isn't an media player, it's a GUI."

Themes are the UI. When you customize your theme to put buttons where you need them, then you're making your UI more useful to you. I thought customization was a big whoop-dee-doo feature of Linux. Now it's being poo-poo'd?

Competition...

[zzxc]

Competition in open source projects is mostly a good thing. This new GUI may make X11 developers improve to keep up. However, different projects like this also create lack of standards. This may require people to use two GUIs, with different applications running on each one. With Windows, every version retains legacy compatability for almost all applications written for a previous version. However, this becoming popular would make it required to run two GUIs to run all Linux applications. Rather than expecting developers to conform X11 emulation should be implemented.

On alternate graphics layers.

[Anonymous Coward]

Alternate graphics layers have been around for a long time. Some of them have significant advantages over X11.

So far the only one which has really gained prominence is the frame buffer device that most modern Linux distributions use when booting. There is even a port of QT to that, and it is sometimes used as the only graphics device in embedded platforms. It has the great advantage of being really lightweight, but it is probably even slower and much less featureful than X11.

Another one is the Y server, which was used in some PDA's until public outcry over lack of source compatibility forced the manufacturer to put in X11 instead (remember that, Slashdot?). Before that there was also svgalib. I don't think anyone cried over that going away.

The issue is support--there are tons of toolkits and applications available for X11, and the networking features are neat and useful once in a while (very often for some people, including myself). Others start with a base of pretty much nothing. That means that it is really hard for them to gain acceptance, even if they are superior from viewpoints such as being smaller, faster, and easier to program.

I personally think that we are going to be stuck with all the cruft and slowness of X11 for a very long time.

Re:On alternate graphics layers.

[Fnord]

The one other issue I've heard of is that since acceleration is implemented in userspace, the server can't block on a hardware interupt and so ends up doing a bit more busy waiting than it should. Though this is only an issue on a machine with a pegged cpu.

Re:On alternate graphics layers.

[Cyno]

What is slow about X?

It plays my OpenGL games just fine. Plays my movies and TV shows just fine. Flips between virtual desks faster than I can blink. What's this slowness everyone keeps talking about?

Maybe everyone buys the same video cards or something, I don't know. I don't have any complaints.

Re:something i always wondered about

[tuffy]

why is linux so beholden to X?

Because when Linux was written in the early 90s, most Unix systems were using X11 for a GUI and that's where most of the open sourced apps were. Inventing a whole new graphic system would've been a huge waste of effort. Arguably, it still is. But for embedded systems, an alternative isn't a bad idea.

Milestone

[bigjocker]

I'm pretty sure that this will set a milestone in the Linux development history.

We have multiple distributions (SUSE, RH, MDK), multiple WM (Gnome, KDE, E), multiple Office Suites (KOffice, OpenOffice, AbiWord), imaging software, network tools ... even multiple kernels where to choose from, but we are stuck with only one graphical environment.

I know there has been a lot of advancemente in the FB handling, but officially, if you want to have some windows you need X.

Personally, I love X, but I for one can see (and have seen a lot of) people complaining about X; and from efforts like this one only good things can happen.

Re:And the .iso mirrors are ?

[jdray]

Was anyone whining? Maybe my filter (1+) is set too high, but I didn't hear any whining. In a commonly-open source community, freely downloadable ISOs are, well, common. A fair mistake, I'd say.

Having said that, I agree that the base expectation of things being free is somewhat overused. And, if it's a good implementation (I can't tell a thing, 'cause the site's slashdotted), I wouldn't mind paying $40 for it.

what's the value proposition

[wfmcwalter]

IMNAMD (I am not a marketing dweeb) but I have to ask - what's their (Rocklyte's) value proposition?

There are several free, reasonably mature windowing environments available for linux already, many featuring hardware acceleration. Several are suitable for embedded use. Why do I want to spend $40 for this? (I'm not being rhetorical - the site isn't accessable). There are innumerable linux distributions, several of which boot straight from CD without install.

Frankly, the speed differentials Scitech quotes (over Xfree) aren't really all that impressive for most graphics adaptors. Sure, there's a big difference between unaccelerated (e.g. vesa) access and accelerated, but a 20% differential between the 2D performance of one accelerated solution and another just isn't that compelling. Now many applications are _that_ dependant on 2D performance? If I'm that 2D bound I can spend that $40 and get vastly better graphics performance by buying a better card.

The "foo is old fashioned", "foo is too complicated", "foo is SO last century" claims that some make (I dunno if these guys do, as their site is still down) aren't value propositions. Is something significantly faster? Significantly smaller? Significantly more useful features? Significantly cheaper? Those are.

Parenthetically, note that I don't apply this standard to free projects. Someone can go code a new OS just for their own pleasure, and doesn't have to pass a customer-value-proposition test. Why? Cos they don't have customers, and so they're not obligated to provide value to anyone.

why use this over X?

[MikeFM]

I've used X for many years and it still works well for me. IMO it is far better than Windows and is better than MacOS. What benefit does this new gui (or any of the others) have over using X? Themes are no big deal. If one feels like it they can theme just about anything in X. X boots from cd too. The page seems to be /.'d so I can't dig to deep.

For any opposing GUI to make ground I'd say it'll need all the features of X and a compatibility layer to let X apps run on them. At the minimum they'd need to make something like a wxWindows port for their gui.

Replacing X is worse than pointless

[njdj]

People bad-mouth X because their PC happens not to use its power. They complain about its "bloat", because they see it taking 10MB of their 256-MB machine's precious RAM, most of which is idle. They complain about it being "slow", which tells me that they have nothing better to do than play video games.

X is many times better than anything else in the marketplace; X is many years ahead of anything that Microsoft offers; it may be old, but so what? The Internet is old. Is that a reason to ditch the internet?

There is value in having alternatives. For mobile phones, the power of X is not needed and something lighter might be appropriate. But to all those who persist in bad-mouthing X, I say: look beyond what 's good enough for the PC in your bedroom right now. Find out what X is really about. It's still leading-edge and is one of the advantages Linux has over its competitors. Does it need improvement? Of course, like pretty much everything that's used. But it's the best base we've got for building on. Discarding it and going back to a Microsoft-like GUI would be a giant leap backwards.

Re:Replacing X is worse than pointless

[Theodore Logan]

> They complain about its "bloat", because they see it taking 10MB of their 256-MB machine's precious RAM, most of which is idle.

This is certainly not a worst case scenario, and you know it. That you adress the problem in this way furthermore attests that you don't know how X works. The answer isn't "blah! You don't need that RAM anyway" but "it doesn't use more RAM!" The reason it appears as if that's the case is that much of the memory it consumes is actually used by applications running on it, since X stores their pixmaps. Taken together X doesn't use more memory than other systems.

> They complain about it being "slow", which tells me that they have nothing better to do than play video games.

So, your counter argument here is that if people think X doesn't suit their needs, they have the wrong needs? I.e. there's nothing wrong with X, only with the people who use it. This is, as I'm sure you understand without me having to point it out to you, an inappropriate way of looking at things. By definition, X is bad if it doesn't suit the needs of its users. If the users want to play games, and X can't handle that, X needs to be rewritten or avoided rather than users must be trained to stop enjoying games.

Besides, many other apps have the same problems with X as games have. Still, the proper way to address this question is to say "X isn't really that slow. XFree86, on the other hand, is, and even that is getting better by every release." Again, you have no idea what you're talking about.

> X is many times better than anything else in the marketplace

Better at what? You don't say. Yet you have just, yourself, provided a lot of reasons for the contrary position.

> X is many years ahead of anything that Microsoft offers

You wouldn't know.

> it may be old, but so what? The Internet is old. Is that a reason to ditch the internet?

Ridiculous comparision. The Internet is not a piece of software. Besides, nobody wants to ditch X merely because it's old - they want to ditch it because they think it sucks. The talk about it being old is merely an excuse for it's suckiness.

> look beyond what 's good enough for the PC in your bedroom right now.

Why should I do that? That's precisely what I, and everybody else for that matter, care about.

> Find out what X is really about.

What is it about, why don't you tell me? According to you, it's not about suiting it's users' needs, and it's not about being good for their computers. Curiously, however, this is what people refer to when they say "X is bad." What you're saying is only that "yes, it's bad in the way everybody thinks it is, but that's not the way that matters. It's good in a lot of other ways, that people don't care about. And this is a very good reason for keeping it even for the people and purposes it doesn't fit." I understand that I'm repeating myself here, but all your "arguments" seem to be answerable in the same way. It's also amusing that, all your ranting notwithstanding, you haven't provided a single example of something X does good, but a lot of examples of things it does poorly.

> It's still leading-edge and is one of the advantages Linux has over its competitors.

Everyone tech-savvy person I know who still use Windows do it because they need it to play games, something not easily done with X even for the few games that exist and are optimized for it. How exactly is this an advantage for Linux over its competitors?

> Does it need improvement? Of course, like pretty much everything that's used. But it's the best base we've got for building on.

You're missing the entire point. The arguments for ditching X is that its architecture simply isn't a very good one for the purposes it's now being used for. If anything at all is wrong with X, it is precisely it's base.

Still, I don't think we should ditch it. Yet, knowing nothing about it and just reading your post I would arrive at the opposite conclusion. You're just a zealot, obviously embracing Linux and X more to fit in with the Slashdot crowd than because you understand their real advantages. It's embarrassing and you are doing the community a disservice.

Re:XFree86 good, not bad

[aussersterne]

Complaints about the slowness of X in Linux are also bogus and are down almost entirely to drivers. Many drivers that have been written by the OSS community tend to have been written essentially by reverse engineering the hardware, without manufacturer support. As a result, they often suck.

Drivers which have been provided by the actual manufacturers of the graphics hardware (as is the case in the Windows world) fare much better.

A perfect example is NVidia hardware, because both free and manufacturer provided solutions exist. The nv driver included with XFree86 is fairly slow in 2d and provides no 3d support. On the other hand, if you download NVidia's Linux drivers for XFree86, you get mind-numbing 2D acceleration and blazing fast 3D acceleration at the same speeds as the Windows drivers, with full OpenGL support.

Unfortunately, because the NVidia drivers aren't OSS, most distributions don't install them. Users install Linux, get sub-par graphics performance, and decide that "1) Linux graphics are slow, 2) X provides Linux graphics, 3) ergo, X is slow" and never even realize that they could increase the throughput of their graphics subsystem manyfold simply by downloading a better driver.

It's really an issue all across the Linux world -- poor driver support because of uncooperative manufacturers... it's just than in X, a poor hack of a driver is much more obvious because the user interacts with it directly.

Re:something i always wondered about

[eht]

Of course I'd rather have true freedom of choice and if I want to run a propritary OS and software I should have that right too, I'd rather have choice than freedom.

Non-X GUIs: What's Really Needed

[istartedi]

What's really needed in a non-X GUI (in fact, all GUIs) is support for higher level APIs so we don't have to care about the underlying GUI. That begs the question, what do developers use most often?

I'd be willing to wager that there is a large percentage of Windows software that uses the GUI's APIs directly--Win32 or one of the popular wrappers like MFC or OWL. On *NIX, GTK is probably the most popular.

There are high-level wrappers that will allow you to target Win32 and *NIX with just a recompile. wxWindows leaps to mind. However, I wager that the percentage of people using them is small, although the following is growing (doesn't AbiWord use wxWindows?).

Given that, I'd probably want to see GTK and wxWindows apps running on top of a non-X GUI before I'd use it. A Win32 subset would be sweet. No, not Wine. I don't want to swallow an elephant just to get a peanut. Full Windows emulation is overkill. I would just like to have Win32 API functions so I could recompile apps that use the APIs directly. I (and thousands of others) have written our own Win32 wrappers. For alternatives to succeed, they need to be able to pull in as much software as possible.

Oh crap... I can't even check the website to see what higher level APIs it supports. D#$% /.

Re:famous AmigaOS GUI?

[nexusone]

That is only your opinion, I found the Amiga's GUI easy to use.
Program did not crash all the time vs my windows machine.
The Amiga had some of the coolist and very creative programs years ahead of windows.

The Amiga died because everyone had IBM clones in schools and work, thus this is what people got for their homes.

Also the US Management at the time did not premote the computer system well.

There is the funny story about how Bill Gates was showing off his 95 multimedia computer, with the computer graphics for it being generated by an Amiga computer hid under a desk.

$40 for this?

[dh003i]

Ok, so they've got this microgui environment. It's not significantly faster than X (20%, according to *them*). It can't run nearly as many apps as X. Oh yea, it offers *nothing* that isn't already offered by a solution that is free as in beer and free as in freedom (PicoGUI). So, why exactly should I waste $40 on this piece of shit, when it isn't that much faster than X, can't run many programs, probably isn't going to be the future of the FOSS community, and offers nothing that isn't offered by Pico's FOSS solution?

There is no reason.

As for performance, the $40 you spend on this crap could be spent getting a better graphics card. I believe you can get a GeForce 2's now for $30-$50. That's what I use on my current Gentoo GNU/Linux system (with WindowMaker). Guess what, no performance problems -- at all [qualifications: 1GHz CPU, 256MB SDRAM, 7200rpm ATA-100 HD].

People are really stupid when it comes to buying the latest greatest whatever. Here are the specs on my current PC:

1.1GHz AMD T-Bird CPU
256MB SDRAM
60GB 7200rpm ATA-100 hard drive
64MB GeForce2 GTS
19" Monitor

Guess what? It was fast when I bought it (a year ago). And guess what? It's still fast. It has not magically been transformed into a lumbering beast. For the home user, there is really no reason at all to buy anything other than that which sits at the best performance/price ratio. You can get an excellent system today for under $1000 that will be able to do anything you want as far as productivity goes, and will play most all games just fine.

My general plan -- and I think it's a good one -- is to upgrade once something 10 times better than what I have is available at an optimal price/performance ratio. And that's only if I have some need.

It's amazing to me how many home users are tricked into believing that a 2GHz CPU is somehow going to make their internet browsing experience any better, or make programs load faster and make Word work better. For almost all typical uses of a computer, you will *never* need to ugrade. You will only need to upgrade if you want to keep up to date with the latest games or if you want to do computationally intense computational work. Even then, you can still do so at the best price to performance ratio.

Re:Myabe X11 just needs another revision

[serbanp]

Sorry but X doesn't work that way. Most of the X packets are related to vector and GC operations. E.g.

XDrawRectangle(mainDisplay, win, drawGC, ltx, lty, width, height);

I believe that most widgets are made from simple geometric figures and not only from plain pixmaps.

Serban

"aging X11" ???

[nurb432]

Bah, X is an established standard that works well.

So few people truely understand what makes X tick is why so many people bash it..

X is wonderful, its the crap that runs on top X that tends to suck and give X a bad name.

Re:famous AmigaOS GUI?

[schon]

The AmigaOS GUI sucked ass and prevented people from buying a technically cool machine.

No, it really didn't.

The AmigaOS GUI was one of the most fluid, easy to understand GUI's I've ever used - it encouraged multitasking, instead of (seemingly) being designed to prevent it (quick - load two Mac/Windows/X word processors, make them full-screen, send one to the back and try to continue using it. Trivial on the Amiga.)

It's one of the things I miss most about it.

Funny thing about that is...

[Paradox]

Having used both NeXT and OSX machines recently, I think this "dumbed down" statement is funny considering how much more feature-rich OSX is compared to NeXT.

I guess it's true you just can't please some people.

Re:something i always wondered about

[Anonymous Coward]

You could always consider the fact that there are more window managers than KDE and Gnome.

icewm, blackbox, and others allow you to run many KDE/Gnome apps without all the added pain that those environments put you through.

I'm sure Gnome/KDE are wonderful for the novice, so I'm not bashing them, but for the 'power-linux-user' who wants speed and small startup times, they are not the best.

Problems with X

[binaryfeed]

Okay. I've read so many damn threads on this topic that I decided it's time I post my own opinion. Facts about X speed:

• X is not slow
• Some X video drivers are slow.
• The slowdown caused by Network transparency is negligible.
• KDE and GNOME are piggishly slow. I use both because I'm willing to sacrifice speed for functionality.

Facts about X usability:

• Configuration is difficult, even for experienced users.
• Cut-and-paste style should be configurable.

What X needs:

• A way to send less data over the wire for toolkits such as QT / GTK+.
• Easier configuration and setup.
• Pluggable cut-and-paste architecture that can be more easily used by the other toolkits.
• Better video drivers*.

* I know ... we're out of luck here without help from the video card manufacturers.