User: Password:
|
|
Subscribe / Log in / New account

Distribution quotes of the week

Wayland's not a usable default yet. It'll probably be packaged in F15 as something you can play with. We don't even have a complete list of transition criteria yet, let alone a timeframe for switching the default. But it's likely to happen eventually because it's a serious win for a lot of things, and the downsides are pretty negligible despite the fear from the peanut gallery.
-- Adam Jackson

The only hardened opinion I have is that network transparency is an essential opinion. Beyond that I have no clue. I'm waiting to be educated. If only Adam Jackson were responding I would have walked away satisfied by now. Perhaps I should ignore everyone else. :)
-- Gregory Maxwell
(Log in to post comments)

Distribution quotes of the week

Posted Nov 12, 2010 20:18 UTC (Fri) by daglwn (guest, #65432) [Link]

Gregory is absolutely right. The communication WRT Wayland has been terrible. No one can give a straight answer about network transparency. I don't care if Wayland supplies it or some other layer supplies it. I just care that it's available on day one.

Distribution quotes of the week

Posted Nov 12, 2010 20:43 UTC (Fri) by jspaleta (subscriber, #50639) [Link]

There is a lesson to be learned here. When luminaries who aren't deeply involved with a project start talking about it out of the blue, people should not just assume that particular person knows what he is talking about it. Premature notoriety can work against a project.

Discussions about technology adoption for disruptive technologies really need to be lead by people involved in the development of those technologies.... not by enthusiasts...and certainly not by millionaire dilettantes who is are getting bored with the day in and day out tedium of funding an unprofitable business venture and want to stir up trouble in other projects they aren't currently involved in the development of.

-jef

Distribution quotes of the week

Posted Nov 12, 2010 22:33 UTC (Fri) by daglwn (guest, #65432) [Link]

Jef, your obsession with Shuttleworth/Canonical is really getting tiring. I do not want this thread to turn into a referendum on Ubuntu. Please, just let it go.

Distribution quotes of the week

Posted Nov 12, 2010 23:21 UTC (Fri) by jspaleta (subscriber, #50639) [Link]

I speaking directly to the comment that the communication with regard to this project has been horrible. Lest you forget that this project was flying pretty much under the radar and just plugging along until one particular very high profile person decided to talk about it and then everyone else took notice. If that one high profile external hadn't mentioned it would we be talking about it at all now? Probably not.

Did the external person's comments take the project participants by surprise as much as it surprised all the other externals? Were they given a heads-up in order to be ready to answer technical questions as part of the conversation they did not start which was happening external to the project's normal communication channels?

So pre-outing by the external, what was this project really suppose to do to communicate better? I'm not sure there is much. This project just firmed up into a distinct freedesktop project, before that it was a collection of code cattered around personal repositories.

-jef

Distribution quotes of the week

Posted Nov 14, 2010 22:24 UTC (Sun) by nix (subscriber, #2304) [Link]

I'd run a poll regarding whether Jef is more obsessed with Canonical than its own employees, but the answer would so clearly be 'yes' that running it would be a waste of time.

Distribution quotes of the week

Posted Nov 14, 2010 22:14 UTC (Sun) by nix (subscriber, #2304) [Link]

I also care that it works at least as well as it does now: i.e. having a 'network transparency' layer that only works if apps are written to take advantage of it is useless, because none of them will. (Witness the 'keep roundtrips down!' X mantra and how well *that* worked.)

Distribution quotes of the week

Posted Nov 14, 2010 10:39 UTC (Sun) by CChittleborough (subscriber, #60775) [Link]

AFAICT, the only substantial user-level documentation the Wayland project has produced so far is this overview. It merits careful reading.

Distribution quotes of the week

Posted Nov 14, 2010 15:16 UTC (Sun) by mfedyk (guest, #55303) [Link]

ok, move the compositor into the xserver process (as a plugin) and you're done. no more wayland.

Distribution quotes of the week

Posted Nov 14, 2010 20:57 UTC (Sun) by quotemstr (subscriber, #45331) [Link]

But that's not acceptable to the Wayland crowd because it won't kill the X11 protocol. After all, it's stuffing bytes over a unix-domain socket that's responsible for the lackluster Linux desktop, not terrible typography or a disdain for usability. Nope. Definitely that unix-domain socket.

Distribution quotes of the week

Posted Nov 15, 2010 1:21 UTC (Mon) by mfedyk (guest, #55303) [Link]

don't local x clients use shared memory with xshm? then it's not even transported over unix domain sockets.

Distribution quotes of the week

Posted Nov 15, 2010 22:29 UTC (Mon) by pflugstad (subscriber, #224) [Link]

I think the problem is all the round trips kill performance, even over shared memory. If you get rid of X11 protcol, you can get rid of all that.

Distribution quotes of the week

Posted Nov 15, 2010 23:12 UTC (Mon) by quotemstr (subscriber, #45331) [Link]

In the local case, "round trip" reduces to "context switch". Where is the evidence that Wayland requires fewer context switches than Xshm? Even if there *is* a slight reduction in the number of context switches, I doubt the difference is important. Linux context switches are very fast and don't even come close to being a performance bottleneck.

Distribution quotes of the week

Posted Nov 16, 2010 8:02 UTC (Tue) by mfedyk (guest, #55303) [Link]

it seems a lot would be solved by using xcb's async api. using xlib or xcb's xlib wrapper only allows one operation in flight at a time. that, I believe us a major source of issues people are seeing.

Distribution quotes of the week

Posted Nov 18, 2010 13:30 UTC (Thu) by renox (subscriber, #23785) [Link]

Sssh, you're messing with the play, the obvious guilty part when there's a GUI issue *must be* X!

Let's ignore that when an issue exist (Xlib doesn't play well with multithreading) and is fixed (XCB) then even when toolkits don't use XCB, it *must* be X's fault.

Distribution quotes of the week

Posted Nov 17, 2010 20:28 UTC (Wed) by oak (guest, #2786) [Link]

Some other X problems are:

* Security

* Clients being able to heavily affect X memory consumption by using pixmaps (so that the new kernel methods for controlling memory usage are much less efficient)

Does Wayland "fix" those issues too?

Distribution quotes of the week

Posted Nov 27, 2010 21:38 UTC (Sat) by job (guest, #670) [Link]

X security is complex beast which is not quite what you want. But in practice with SSH it's all automatic and that's what people use nowadays.

Distribution quotes of the week

Posted Nov 29, 2010 11:50 UTC (Mon) by mfedyk (guest, #55303) [Link]

> * Clients being able to heavily affect X memory consumption by using pixmaps (so that the new kernel methods for controlling memory usage are much less efficient)

Aren't most reports of this just memory that's mapped into GPU memory?


Copyright © 2010, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds