X.org 7.4 released [LWN.net]

X.org 7.4 released

Posted Sep 25, 2008 7:55 UTC (Thu) by alankila (subscriber, #47141) [Link]

Ok, this is a rant.

For Apple & MS it must be said that most users do not need to send a bug report in the first place. My experience has been that things generally work on Win32 land. Although I haven't got any hands-on experience with OS X personally, the users over there at least are pretty vocal about the "things just work" part. If true, it might explain why many previously Linux-using hackers are switching to OS X as well.

Here is the radical suggestion: maybe we should stop changing stuff around so much and instead try to make a decent software ecosystem with development at the tip of the free software tree, and for instance keep both free and proprietary stuff working together in harmony?

I have my own cow in the ditch here as I'm on an ATI X1250 system whose 3D driver is again broken for last 3 months because of some technically brilliant but almost certainly equally unnecessary change in where some server private data is stored on Xorg 1.5.

Here's the problem: they knew it wouldn't work with fglrx but they released that incompatible change anyway, which was equally certainly packaged in a distribution and pushed to users. Just great, and reminds me of a very similar story with Pulseaudio. This attitude of accelerating development by treating users as a pool of people for which it doesn't matter if their shit breaks is part of the problem, and its essential consequence is lack of users and thus lack of world domination.

Lesson is: if nobody is testing it, then nobody is wanting it. Developer can only sell his new stuff by making it BETTER than the old stuff. It's not complicated equation, but as users seem to have no intrinsic value in this ecosystem (apart from being unwilling beta testers), their concerns are consequently irrelevant.

I think it drives the few users that do stick along away. After experiencing the crown jewels of the open source world over a multi-year period, the users treated this way will happily embrace closed systems like OS X or Win32 because the companies making these things actually need to keep their users reasonably happy.

I think this seriously impedes Linux's march towards world domination. To win the game of numbers, Linux needs _especially_ the users that aren't contributors, simply because there are so many of them. It's time open source developers looked at users and spent time thinking what is important for them. (And if this drives the developers away who don't care about users, I'd be happy about it.)

X.org 7.4 released

Posted Sep 25, 2008 10:04 UTC (Thu) by modernjazz (subscriber, #4185) [Link]

In a sense I agree with you, in that I have learned to approach system
upgrades with a fair amount of trepidation---bugs/inconveniences that
I've learned to mostly work around might be fixed, but there are usually
new problems that I have to learn to cope with. For someone who's trying
to get work done (rather than mostly have fun with the computer), that
does become a bit tiresome. I do report bugs, and it's delightful to see
just how often they get fixed (the system does work!), but I also can
agree that in an even better world I wouldn't have to report bugs because
there wouldn't be any to report :-).

I haven't seriously used the proprietary systems in ages, but I have the
impression that they are not free from such problems, either. It seems
that Apple must have it easier because of the limited range of hardware
(presumably they really can test an upgrade on every recent system),
whereas Windows and Linux have a much more diverse set of hardware to
cope with.

However, by and large I've basically resigned myself to this state of
affairs. In free software land, there are too many truly important things
to do (and cleaning up X must rank near the very top) for my minor
inconveniences to become an argument to slow down development. After all,
I develop a free application myself (nothing at the "core" of our systems
like the kernel or X), and I'm aware of the ways in which freedom to
tinker will bring long-term benefits to the users.

My view/hope is that the "worst" of this turmoil will be somewhat
transient: the truly fundamental changes in X will probably continue for
another several years (even though some of the biggest, baddest seem like
they will be fixed sooner than that), but several other key domains
(e.g., wireless, suspend/resume, and maybe even sound) seem like they
could "settle down" in the relatively near future.

It's not to say that there will come a point at which development could
stop (far from it), but it does seem that developers are taking the
opportunity to tackle a lot of really important problems right now. More
power to them. And I'll just keep crossing my fingers when I upgrade my
system.

X.org 7.4 released

Posted Sep 25, 2008 19:42 UTC (Thu) by alankila (subscriber, #47141) [Link]

Believe it or not, I also tend to assume that the problems are essentially transient. For instance, I suppose I can appreciate the fact that we only need to get Pulseaudio working once; after it flushes the bugs out of the audio system we should all benefit from having an audio stack which is smoke tested by its unusually complicated requirements.

It's just hard to keep that in mind each time that the audio stutters now during CPU contention (on a dual core system, no less!) and in the new world order flash works maybe 10 % of the time (installing libflashsupport makes the browser freeze), so every youtube video takes 10 reloads to display if I'm lucky. I suppose I can live without flash for now, but it makes me truly long for an OS where things just work.

(And sometimes pulse and all pulse clients jam, presumably because of some hitherto unsolved ALSA bug, and the only solution is to kill everything, including Pulse, and start anew. An audio server should be programmed like a tank, but I guess it's vulnerable to hardware driver bugs like everything else. The fact it's unable to recover without kill & restart must count against it somehow, though.)

X.org 7.4 released

Posted Sep 29, 2008 15:03 UTC (Mon) by whitemice (guest, #3748) [Link]

>and in the new world order flash works maybe 10 % of the time

Man, I hear this allot. But my flash always works and has always worked for years now. You-Tube never causes any problems. I must have already arrived in the "new world".

Does openSUSE just rock that much?

X.org 7.4 released

Posted Sep 29, 2008 15:06 UTC (Mon) by whitemice (guest, #3748) [Link]

>My view/hope is that the "worst" of this turmoil will be somewhat
>transient: the truly fundamental changes in X will probably continue for
>another several years (even though some of the biggest, baddest seem like
>they will be fixed sooner than that), but several other key domains

Agree. The LINUX desktop has made huge, yet quite, gains in the past couple of years. Usability is way ahead of there it was. The great majority of things do "just work".

>(e.g., wireless, suspend/resume, and maybe even sound) seem like they
>could "settle down" in the relatively near future.

Agree, the wireless is improved, but needs to go a bit further. Suspend/Resume is still pretty rough - but that is tough as every vendor or even model nuances the configuration. Sound works great for me after a few updates from the openSUSE repository, it has stuttered or crashed in over a month now.

Things just work

Posted Sep 25, 2008 13:35 UTC (Thu) by Thalience (subscriber, #4217) [Link]

"Things just work in Win32 land"

This is largely true when you buy a pre-configured system from Dell or other integrator. It stops being true as soon as you change the system in any way. Even just re-installing the system from the integrator's install media can be a challenge for the non-technical user.

Apple does a much better job on this than MS, but the experience is still far from perfect. Just like any other system, some upgrades introduce regressions. Especially when third-party software and hardware is involved.

X.org 7.4 released

Posted Sep 25, 2008 14:23 UTC (Thu) by nix (subscriber, #2304) [Link]

I have my own cow in the ditch here as I'm on an ATI X1250 system whose 3D driver is again broken for last 3 months

Whose *binary-only* 3D driver is broken.

Why anyone uses fglrx voluntarily is completely beyond me. Even when it's working it seems to specialize in display corruption and lockups.

X.org 7.4 released

Posted Sep 25, 2008 19:08 UTC (Thu) by alankila (subscriber, #47141) [Link]

I like compiz's screen zoom feature. It is nice way to rest eyes when you are dealing with text in only small screen area.

Also I have in mind to develope something that creates OpenGL texture through CPU, and it will require the texture-from-pixmap extension to work acceptably fast. (So that there is no copy/format translation when the texture leaves CPU and enters the GPU.)

In other words, I need/want some 3D features. The open source radeonhd driver does not currently provide them. On top of that, the 2D is also pretty slow -- maybe completely unaccelerated -- as it looks like X spends a lot of time blitting screen areas.

X.org 7.4 released

Posted Sep 25, 2008 21:17 UTC (Thu) by nix (subscriber, #2304) [Link]

Ah, right. As a longtime fvwm user I've become resigned to never ever
getting any of these nifty new transparency-et-al features, so I tend to
forget that for other people they're a fact of life :)

X.org 7.4 released

Posted Sep 25, 2008 14:43 UTC (Thu) by zlynx (subscriber, #2285) [Link]

"For Apple & MS it must be said that most users do not need to send a bug report in the first place"

It should be said that's its nearly impossible to send a bug report to either of these companies if you're an ordinary user.

For example, since Vista was released I've reported six bugs to MS. Well, 3 were probably feature requests.

I didn't want to use a forum or MS Connect or call MS. I wanted something simple so I dug around Vista's menus and help system. All I could find was a general "feedback" form and I used that.

I only know about MS Connect because I had done some beta stuff. I have no idea how regular MS users find out about it.

And the only way I know to request features / bug fixes from Apple is to use IM to tell a friend who happens to work there about it. :)

X.org 7.4 released

Posted Sep 25, 2008 19:41 UTC (Thu) by man_ls (subscriber, #15091) [Link]

And the only way I know to request features / bug fixes from Apple is to use IM to tell a friend who happens to work there about it. :)

It is easy: just join ADC, get free betas and pay for the privilege of reporting bugs to Apple.

X.org 7.4 released

Posted Sep 25, 2008 20:16 UTC (Thu) by foom (subscriber, #14868) [Link]

Anybody can register for a free ADC account, which then allows you to report bugs to apple. Doing so can be a bit depressing: it feels like communicating with a black hole...you likely won't get any status information whatsoever, even a "bug confirmed, we're working on it" type of notification, until the bug fix has actually been released into a new version of the software.

But it does actually work: engineers working on the software do see your bug reports, and will act on them.

Now, the only way to get status information on bugs you care about is indeed to IM a friend who works there. :)