[go: up one dir, main page]
|
|
Log in / Subscribe / Register

Wayland - Beyond X (The H)

Wayland - Beyond X (The H)

Posted Feb 13, 2012 21:29 UTC (Mon) by dlang (guest, #313)
In reply to: Wayland - Beyond X (The H) by JohnMorris
Parent article: Wayland - Beyond X (The H)

No, depending on who you ask network transparency is either a feature that's not needed in today's world, or it is something that they will look at adding to Wayland later on (i.e. submit the patches)

There is some talk that you may be able to run a X server inside Wayland.

overall it doesn't look good in terms of the network access as far as I can see.

to post comments

Wayland - Beyond X (The H)

Posted Feb 13, 2012 21:48 UTC (Mon) by Company (guest, #57006) [Link] (30 responses)

I think the common answer is "VNC is better than X anyway."
And VNC will work fine with Wayland.

Wayland - Beyond X (The H)

Posted Feb 13, 2012 21:59 UTC (Mon) by valyala (guest, #41196) [Link] (12 responses)

VNC eats a lot of bandwidth. NX is much better in this respect - http://en.wikipedia.org/wiki/NX_technology . But it looks like NX won't work well with Wayland.

NX dead end?

Posted Feb 14, 2012 8:17 UTC (Tue) by amtota (guest, #4012) [Link] (11 responses)

NX does seamless windows which is great, but it is also closed source nowadays (v4.x onwards), and although v3.x is open-source it is not longer properly maintained.
Try http://xpra.org/ (which I maintain and does seamless windows and more)

NX dead end?

Posted Feb 14, 2012 9:09 UTC (Tue) by sitaram (guest, #5959) [Link] (7 responses)

How old is this project? Why doesn't it get more attention? Why haven't I heard about it till now? Why Why Why???

In one shot you solve one of my biggest problems (long story...) and I only find out because I happened to read all the *comments* on a story?

How old is this project? Why doesn't it get more attention? etc

Posted Feb 14, 2012 10:09 UTC (Tue) by amtota (guest, #4012) [Link] (5 responses)

> How old is this project?
The original version (this is a fork) was started sometime in 2008 by Nathaniel Smith, the earliest reference I can find on LWN is http://lwn.net/Articles/276855/ - "Announcing xpra v0.0.4" - April 2008. I've spent the last 3 years actively improving/packaging/supporting it.

> Why doesn't it get more attention? Why haven't I heard about it till now? Why Why Why???
Because I spend my time trying to make it work better (coding mostly), not writing press releases? Although I should probably do a bit more of the latter..

> In one shot you solve one of my biggest problems (long story...) and I only find out because I happened to read all the *comments* on a story?
Glad it helped.

How old is this project? Why doesn't it get more attention? etc

Posted Feb 14, 2012 11:34 UTC (Tue) by Cato (guest, #7643) [Link] (4 responses)

Sounds really useful, particularly since you have a Windows port and I often need Windows to Linux interoperability.

Please send announcements of new releases to LWN and Freshmeat (now Freecode) at least - that would get you a lot more visibility and not take much time.

It would also help if you mentioned on the site the key words 'remote access' and that you cover a similar problem space to NX (and to some extent VNC) - it's not clear from the xpra site that it works well over low bandwidth connections. Also useful if you mention that xpra over SSH is supported - it's given as an example but mentioning it as a feature would be good.

How old is this project? Why doesn't it get more attention? etc

Posted Feb 14, 2012 14:02 UTC (Tue) by amtota (guest, #4012) [Link] (3 responses)

> Please send announcements of new releases to LWN and Freshmeat (...)
Will do for LWN, problem is that I tend to do one or two releases a month and don't really want to spam LWN.. The original (unmaintained) version is on freshmeat so unless I rename this fork, I can't update that.

> (..) mentioned on the site the key words 'remote access'
> Also useful if you mention that xpra over SSH is supported
Both done, also added more info to the home page. Thanks!

> it's not clear from the xpra site that it works well over low bandwidth connections
It's OK, I've added some info to the site. FYI: on localhost and using mmap, it's seriously fast.
By design NX will remain king for certain applications in this department, but this advantage is quickly disappearing for the exact same reasons that work in favour of Wayland (newer apps rely on toolkits rather than the old and crufty X11 APIs)

(sorry for hijacking the thread)

How old is this project? Why doesn't it get more attention? etc

Posted Feb 14, 2012 19:23 UTC (Tue) by dlang (guest, #313) [Link]

I don't see an update or two a month as being a big problem for blurbs to LWN

and it sure would help visibility of the project.

How old is this project? Why doesn't it get more attention? etc

Posted Feb 17, 2012 21:32 UTC (Fri) by wookey (guest, #5501) [Link] (1 responses)

wow, screen for X, and I knew nothing about it. Absolutely no need to apologise for 'thread hijacking'. This is a very useful thing indeed, which apprently not enough people know about. I've waited for NX for many years but it's always been a big pain in practice (forks, licencing, server not in Debian). Anything that gives me reconnectable-at-home/work X apps is _extremely_ useful. X over ssh is OK, but it doesn't give you the 'reconnect' feature.

If it's good it might even get me out of a range of terminal-based core practices which are text-based because the 'reconnectable single instance' aspect is the most important thing (irc, email).

How old is this project? Why doesn't it get more attention? etc

Posted Feb 17, 2012 23:38 UTC (Fri) by khc (guest, #45209) [Link]

I've been using winswitch (http://winswitch.org/) which wraps around Xpra and gives you a pretty GUI to work with.

NX dead end?

Posted Feb 14, 2012 13:44 UTC (Tue) by mjthayer (guest, #39183) [Link]

> In one shot you solve one of my biggest problems (long story...) and I only find out because I happened to read all the *comments* on a story?

Quite often I find that comments on LWN are as valuable as the stories themselves. (Which are excellent of course.)

Xpra

Posted Feb 14, 2012 15:03 UTC (Tue) by bawjaws (guest, #56952) [Link] (1 responses)

Is the 'xpra' available from the Ubuntu repositories your fork or the original. If the original do you have a link detailing the differences?

Xpra

Posted Feb 14, 2012 17:49 UTC (Tue) by amtota (guest, #4012) [Link]

Ubuntu ships the old version, Debian testing has it though so maybe Precise Pangolin will ship it? (Debian is already quite a few versions behind, but it will sync again)
As for the list of changes, they are listed right on the home page: http://xpra.org/

NX dead end?

Posted Feb 18, 2012 12:19 UTC (Sat) by Jonno (guest, #49613) [Link]

Please note that xpra is a "VNC-like protocol" (eg it sends images, not render commands, over the network), and represents pretty much exactly what you eventually can expect to get from Wayland.

It is even implemented in the same general way as remote Wayland are intended to eventually be implemented, eg. it is a "compositing window manager" that don't actually composes the windows but sends them over the network instead.

Wayland - Beyond X (The H)

Posted Feb 13, 2012 22:23 UTC (Mon) by elanthis (guest, #6227) [Link] (16 responses)

That is not the common answer. You are one of the many who keep misinterpreting the (probably poorly phrased) answer: a VNC-like remote display protocol is better than an X11-like remote drawing command protocol. That means that a remoting protocol that focuses solely on the upload of (compressed) image buffers is better than a remoting protocol that focuses on sending rendering commands to the server. It does NOT mean that VNC's many other undesirable qualities are meant to be kept in any future Wayland remoting system.

X11 was designed in the day when apps just draw lines and solid rectangles and used server-side fonts. Here in 2012, apps have smooth gradients and rounded corners rendered with complex series of Cairo compositing commands and interesting effects rendering with intricate GL uses that require lots of buffer uploads and feedback. X11's design no longer reflects how applications are written.

VNC is not at all an ideal protocol to use. All the Wayland folks are saying is that the idea of uploading image buffers (like VNC) will be the future. Things like how VNC's image buffers are crappily compressed or how VNC must export an entire desktop rather than individual windows is not something anybody wants to keep around. Hence, VNC itself is not the answer. People mentioning VNC are talking ONLY about the broad general idea of sending image buffers.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 1:09 UTC (Tue) by nix (subscriber, #2304) [Link] (6 responses)

It seems to me that an ideal protocol would upload images to the remote end only when the local application informs the local server of them (capped at most once per display refresh). Sending images on every single display refresh, even if you already sent the same image in the last refresh, is pointlessly inefficient. That's what VNC does, and that is all that's wrong with remote display protocols as compared to remote command protocols.

i.e., do with the remote protocol what Xft and GlyphSets did for client-side fonts.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 2:59 UTC (Tue) by aliguori (subscriber, #30636) [Link] (5 responses)

VNC only sends updates (1) to the dirty region of the framebuffer (2) at the request of the client.

As it turns out, in real life, the framebuffer changes constantly so the use of a refresh rate for something like VNC is to batch updates. But only dirty portions of the screen change.

Where VNC lacks compared to modern remote display protocols is things like offscreen image buffers and z-ordering.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 11:19 UTC (Tue) by nix (subscriber, #2304) [Link] (3 responses)

The framebuffer changes constantly, but when textual output is being displayed it normally changes to display the same bits of stuff over and over again (or the same bits of stuff over and over again, alpha-composited onto a constant background or a constant background pixmap). If the remoting model used by VNC represented this properly, it would be very low-bandwidth indeed. I hope the remoting model used by Wayland, when it eventually gets one, can do the same.

(Yeah, this won't handle the hard cases where people are playing any game, at least any game by Jeff Minter, or where people are displaying text in dancing fonts atop animated shaded backgrounds, but text display is a lot of what X, or any windowing system, is *for*. Forget xterms -- think IRC clients or web browsers displaying any textual article. GlyphSets *rule* for bandwidth reduction: no content-unaware compression algorithm could get remotely the same size reduction for the same low CPU use. Discarding that would be terrible.

Given that the inventor of GlyphSets is involved in Wayland I suspect that it will not be forgotten.)

Wayland - Beyond X (The H)

Posted Feb 14, 2012 21:53 UTC (Tue) by aliguori (subscriber, #30636) [Link] (2 responses)

>The framebuffer changes constantly, but when textual output is being displayed it normally changes to display the same bits of stuff over and over again (or the same bits of stuff over and over again, alpha-composited onto a constant background or a constant background pixmap). If the remoting model used by VNC represented this properly, it would be very low-bandwidth indeed.

It doesn't. All modern tool kits do their own font rendering (this is what Pango provides).

Wayland - Beyond X (The H)

Posted Feb 14, 2012 22:59 UTC (Tue) by nix (subscriber, #2304) [Link] (1 responses)

Yeah. They do their own font rendering, rendering to glyphs which are then sent to X as a GlyphSet, right?

Wayland - Beyond X (The H)

Posted Feb 16, 2012 0:05 UTC (Thu) by aliguori (subscriber, #30636) [Link]

Oh, I was not aware of the GlyphSets extension. VNC only has a very simple bitmap based protocol that can provide updates to the bitmap at a rectangle granularity.

It has no concept of offscreen bitmaps, alpha blending, or anything like that.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 11:40 UTC (Tue) by Cato (guest, #7643) [Link]

There is a way to do offscreen image buffers with VNC, though it's a bit of a hack. You x11vnc's client-side caching feature, which allocates a huge long screen buffer, and tell the client that the screen is less tall than it appears. See http://en.wikipedia.org/wiki/X11vnc#Client-side_caching

X11VNC lets you VNC to the main X screen as well, while SSVNC is a handy Linux client that does the tunnelling for you.

I find that UltraVNC on Windows works fine as a client to x11vnc, and TightVNC viewer on Linux is OK too.

iSSH on iPhone almost works but is currently rather crashy for VNC. If anyone else can recommend solid VNC-over-SSH clients for iOS I'd really like to know them.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 2:28 UTC (Tue) by wahern (subscriber, #37304) [Link] (1 responses)

You can't do rendering and compositing for dozens or hundreds of applications all on the same machine. The implication is that Wayland doesn't see a future for thin clients. Or perhaps that the future of thin clients will be HTML.

Perhaps someone should start an X server implementation for HTML5 using JavaScript, WebSockets, Canvas, etc.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 4:24 UTC (Tue) by Cyberax (✭ supporter ✭, #52523) [Link]

>You can't do rendering and compositing for dozens or hundreds of applications all on the same machine.

But you can! A decent 3D graphics card can handle many many many windows just fine. You'll be limited by the network bandwidth, RAM or CPU long before you hit 3D card's limits.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 8:14 UTC (Tue) by amtota (guest, #4012) [Link]

> VNC must export an entire desktop rather than individual windows

Try http://xpra.org/ instead.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 8:31 UTC (Tue) by jch (guest, #51929) [Link] (2 responses)

> a VNC-like remote display protocol is better than an X11-like remote drawing command protocol

Note that there's more to remote X than just rendering -- X is pretty good at event distribution (the mouse pointer is only tracked when the client cares about the mouse pointer position) and avoidance of race conditions (the passive grab mechanism).

The latter is an absolute must over laggy links -- it is what allows opening a menu and choosing a menu entry *before* the menu entry appears. The former is probably helpful on asymmetric links (ADSL lines etc.).

--jch

Wayland - Beyond X (The H)

Posted Feb 14, 2012 23:22 UTC (Tue) by mgedmin (guest, #34497) [Link] (1 responses)

OTOH it's not fun if the network connection stalls when a remote app has a server grab. You can't do _anything_ with your machine (except maybe Ctrl+Alt+F1).

Wayland - Beyond X (The H)

Posted Feb 15, 2012 15:40 UTC (Wed) by nix (subscriber, #2304) [Link]

What we really need is an automatic server-side timeout for grabs. If you take an input device grab and don't release it fast enough, you lose the grab and get sent an event about it (older apps can't deal with that event, but often can recover anyway in my experience, maybe taking a click or two to realize that this menu should be closed really). If you take a full server grab and don't release it fast enough, your X connection is dropped and the grab is released.

('Fast enough' is a difficult question, particularly for mouse grabs which are often held for fairly long periods. A customizable interface is probably best, perhaps something like a minute for keyboard and mouse grabs by default and a second for full-blown server grabs. Clients can opt out of this behaviour, but if any do and are not screensavers they can expect much mockery.)

Would this work, or am I babbling insanity?

Wayland - Beyond X (The H)

Posted Feb 14, 2012 8:44 UTC (Tue) by epa (subscriber, #39769) [Link] (2 responses)

Yes, the current situation is a bit silly: if you run a GTK+ or Qt application over the network you're mostly sending rendered images over the wire, hardly making use of X's remote drawing primitives. (Which is a pity in some ways, I kind of liked the old X11 server-side bitmap fonts, but they don't really fit into the modern internationalized, Truetypey world.)

Wayland - Beyond X (The H)

Posted Feb 14, 2012 11:47 UTC (Tue) by nix (subscriber, #2304) [Link] (1 responses)

That's wrong. You send a GlyphSet once, then after that you just tell the server to 'composite that Glyph I sent you earlier from that GlyphSet *here*'.

Without this remote textual display programs (not just xterms: IRC clients, emacs, you name it) would never be able to scroll at tolerable speed.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 14:28 UTC (Tue) by epa (subscriber, #39769) [Link]

Thanks for the correction.

Wayland - Beyond X (The H)

Posted Feb 13, 2012 21:58 UTC (Mon) by tstover (guest, #56283) [Link] (63 responses)

Wayland might actually help the situation. All of the ideas that conflict with the network transparent X11 model might be able to flourish in the wayland world. Ideas regarding the evolution of X then could then continue to support network transparency with less debate.

Currently Wayland can run inside X, and eventually X inside Wayland.

I'm sure there will be some great things to come from a fresh start.

X is mature, healthy, and will be around for a long long time. Something as CRITICAL and WIDELY DEPLOYED as X over the network is not going to disappear just because an alternative focused on faster local rendering materializes.

...and Yes I am typing this in a web browser - and that browser most certainly IS running on a remote LAN host with ssh X11 forwarding. :)

Wayland - Beyond X (The H)

Posted Feb 14, 2012 2:31 UTC (Tue) by flewellyn (subscriber, #5047) [Link] (2 responses)

>Currently Wayland can run inside X, and eventually X inside Wayland.

But, what about Wayland inside X inside Wayland?

Wayland - Beyond X (The H)

Posted Feb 14, 2012 2:58 UTC (Tue) by Kit (guest, #55925) [Link] (1 responses)

Actually, Xorg already works inside Wayland, both rootless and fullscreen. The Wayland screenshots (http://wayland.freedesktop.org/screenshots.html) of X running inside Wayland are actually taken from inside a (at least I presume) X session... so you have Xorg on Wayland on Xorg!

Wayland - Beyond X (The H)

Posted Feb 14, 2012 8:55 UTC (Tue) by iksaif (guest, #54284) [Link]

Some screens of X inside Wayland inside X: http://xf.iksaif.net/dev/wayland/
native Xorg [ Qt Wayland compositor [ xwayland + x clients ] ]

Wayland - Beyond X (The H)

Posted Feb 14, 2012 8:19 UTC (Tue) by khim (subscriber, #9252) [Link] (59 responses)

Something as CRITICAL and WIDELY DEPLOYED as X over the network is not going to disappear just because an alternative focused on faster local rendering materializes.

You may show of the X over network as much as you want but the simple fact: people don't use it. The biggest hint is the fact that the most common usecase from Windows world (I work late in the evening, decide to continue from home via remote access) does not use X at all. It uses NX or some other kludge, but not X. Was it possible to extend X to support this usecase? Probably - but noone cared. This means that world is already split in two: people who use X networking continue to use it and preach it's usability, while "new generation" does not care at all. Similar to the situation with network filesystem in 1990th (when Netware calmed that something as CRITICAL and WIDELY just can not disappear).

I doubt X will disappear completely: old version or it and old versions of program (X-compatible, not Wayland-compatible) will always be available. But at some point new programs will stop using it and will stop supporting it.

Natural progression of life: either you adopt or you disappear. And X networking refuses to adopt.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 11:25 UTC (Tue) by pboddie (guest, #50784) [Link] (1 responses)

Huh? I can't be the only person to use X over SSH, even if I don't do it as often as I might (and certainly, VNC or xpra would be better for certain programs that "need" a GUI but also run for long periods).

And you mean "adapt" not "adopt".

Wayland - Beyond X (The H)

Posted Feb 14, 2012 21:34 UTC (Tue) by lally (guest, #71211) [Link]

I did, but don't anymore. I use SSH for remote login, emacs/tramp for text editing, and a local web browser for our company apps. I suspect that a lot of original uses for remotely-accessing graphical apps are reduced with the prevalence of webapps.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 11:46 UTC (Tue) by nix (subscriber, #2304) [Link] (36 responses)

You may show of the X over network as much as you want but the simple fact: people don't use it.
You keep saying this even though people keep popping up in response to your words and saying 'I use it' over and over again. Since you're ignoring people directly telling you that your assertion is wrong, yet you keep making it, why should we believe anythign else you have to say?

Wayland - Beyond X (The H)

Posted Feb 14, 2012 19:56 UTC (Tue) by khim (subscriber, #9252) [Link] (35 responses)

You keep saying this even though people keep popping up in response to your words and saying 'I use it' over and over again.

Yup. The very same group of few dozens of people raise huge racket again and again.

Since you're ignoring people directly telling you that your assertion is wrong, yet you keep making it, why should we believe anythign else you have to say?

It's your choice. Either you are creating desktop for the masses (and that means 'I use it' arguments are sent straight to /dev/null where they belong), you you continue to "scratch your own itch" till it's raw.

Long, long ago people who wrote the software were more-or-less the same as people who used said software. Back them it made sense to ask your users and do as they say. Today not only developers are different from users, what's worse majority of user are not presented on the forums where developer participate.

Thus you need to use indirect methods. For example you can count number of Macs (which don't support anything besides VNC) and PC laptops (which at least theoretically can support network transparency) on Linux conferences.

This number clearly grows - and that means that people are not bothered by lack of network transparency. For the applications which are designed to be run over network you can continue to use X - it's quite compatible with Wayland.

But for applications designed to be run locally network transparency is just useless abstraction.

I already wrote about this "you don't exist" phenomenon before.

The fact is: most users value "pretty pixels" way, way, WAY above network transparency. I wrote about this, too.

You may think that "serious feature" should not ever be sacrificed for "useless embellishments", but in reality it's the other way around. If your application is ugly then nobody will want it and if your platform encourages creation of ugly applications then your platform will be ignored by Joe Average completely.

The last example which shows that is surprising: it's Micrsoft's WP7. It was designed with very simple, streamlined, fast interface in mind. And people like that. The problem Microsoft faces: without enough blitz in the interface people will not even try to use it.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 20:29 UTC (Tue) by dlang (guest, #313) [Link]

so you are saying that you don't care that people are using this feature, they aren't enough people to matter, so if they can't use your new windowing system, tough luck.

and you wonder why people get upset at pushing this new windowing software as the future of computing.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 21:26 UTC (Tue) by jonabbey (guest, #2736) [Link] (2 responses)

'Thus you need to use indirect methods. For example you can count number of Macs (which don't support anything besides VNC) and PC laptops (which at least theoretically can support network transparency) on Linux conferences.'

That's odd, both of my Mac laptops support X11.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 21:41 UTC (Tue) by raven667 (subscriber, #5198) [Link] (1 responses)

I think the conversation has now passed some sort of event horizon. Macs support X11 in the same fashion Wayland does

Wayland - Beyond X (The H)

Posted Feb 14, 2012 21:45 UTC (Tue) by jonabbey (guest, #2736) [Link]

Yeah, I realized after I posted that he was speaking more of remote accessibility for native Quartz apps and that my counter was a bit off target.

Wayland - Beyond X (The H)

Posted Feb 15, 2012 1:20 UTC (Wed) by daglwn (guest, #65432) [Link] (4 responses)

> and that means 'I use it' arguments are sent straight to /dev/null where
> they belong

Sums up perfectly the disgust people have at the attitude of the Wayland community.

Wayland - Beyond X (The H)

Posted Feb 15, 2012 1:40 UTC (Wed) by rahulsundaram (subscriber, #21946) [Link] (3 responses)

:"Sums up perfectly the disgust people have at the attitude of the Wayland community."

How so? None of the wayland developers have posted here. khim doesn't represent the wayland community any more than you do.

Wayland - Beyond X (The H)

Posted Feb 15, 2012 4:38 UTC (Wed) by daglwn (guest, #65432) [Link] (2 responses)

Somebody better tell him to pipe down, then, because he's giving them a bad reputation.

Given the quotes in the article, I don't know that he's far off from the developers.

Wayland - Beyond X (The H)

Posted Feb 15, 2012 16:30 UTC (Wed) by rahulsundaram (subscriber, #21946) [Link] (1 responses)

Regardless of what you might think of khim's opinions, I don't believe he is giving them a bad reputation at all unless you blame Wayland developers for his opinions which you have no reason to do at all. If you want to know their opinions, ask them

Wayland - Beyond X (The H)

Posted Feb 17, 2012 2:30 UTC (Fri) by jschrod (subscriber, #1646) [Link]

khim *is* giving them a bad reputation.

If they don't want that, maybe, they should start to get involved in the discussion on one of the important, if not the most important, Linux news sites? Life's a bitch and all that...

Wayland - Beyond X (The H)

Posted Feb 15, 2012 18:57 UTC (Wed) by jedidiah (guest, #20319) [Link] (22 responses)

> Yup. The very same group of few dozens of people raise huge racket again and again.

It's not 1990 anymore. The concept of X over the network is no longer as strange as you would like to pretend. The inclusion of a bad implementation of this idea in MacOS is ample demonstration of this. The network GUI concept is fairly commonplace in the corporate Windows world. If anything, people should be improving these features in Linux rather than trying to abandon them.

If the Wayland crowd have their way we will have a perverse situation where Windows is better at rendering it's GUI across the network than Linux is.

MacOS+VNC is just plain painful and is nothing that should be held up as an example of how Wayland might work.

Wayland - Beyond X (The H)

Posted Feb 15, 2012 19:04 UTC (Wed) by khim (subscriber, #9252) [Link] (20 responses)

MacOS+VNC is just plain painful and is nothing that should be held up as an example of how Wayland might work.

Right. But there are undeniable fact: Mac is the only platform which significantly grown market share over last 10 years or so (Windows is losing it, Linux is stagnating).

Everyone agree that MacOS+VNC combo had nothing to do with that: people hates this side of MacOS. But apparently butter-smooth MacOS experience is highly relevant to said growth - and it's direct consequence is this exact MacOS+VNC combo.

Thus I'm not sure I want to demand network transparency (as many here are doing): sure, it's nice feature to have, but if lack of network transparency will be compensated by great local experience it still can be a net win as MacOS shows.

Wayland - Beyond X (The H)

Posted Feb 15, 2012 19:37 UTC (Wed) by jedidiah (guest, #20319) [Link] (19 responses)

The "butter smooth" MacOS experience is mostly mindless hype. It completely evaporates once your requirements are non-trivial.

Want a Mac? Buy yourself a Mac. Don't sabotage Linux.

Wayland - Beyond X (The H)

Posted Feb 15, 2012 20:12 UTC (Wed) by tstover (guest, #56283) [Link] (18 responses)

Indeed.

The growth of OSX is directly related to non-technical customers being forced to find alternatives to windows over the course of its continued odyssey into free-style mass self destruction. To even suggest otherwise requires completely entering into an alternative reality. Most end users don't want to pay an extra 500% for a computer.

This also echoes the rise of Linux for that matter in the 90s with technical users.

(the ipod/itunes societal downfall is an orthogonal issue)

Wayland - Beyond X (The H)

Posted Feb 15, 2012 21:52 UTC (Wed) by khim (subscriber, #9252) [Link] (17 responses)

The growth of OSX is directly related to non-technical customers being forced to find alternatives to windows over the course of its continued odyssey into free-style mass self destruction. To even suggest otherwise requires completely entering into an alternative reality.

And how we've entered realm of pure fantasy and self-delusion.

To even suggest otherwise requires completely entering into an alternative reality.

Perhaps. But then it just means that we all were moved to this alternate reality recently. When you participated last in some developer's conference? Number of Macs dwarfs number of Linux laptops on events like GDC. And not just among PHP developers. Googlers also overwhelmingly prefer them over Linux laptops. And that's in a company which uses Linux almost exclusively on servers and which officially gives supported Linux laptops for the ones who prefer them! If Google engineers are mindless non-technical customers then who are these mythical highly techinical customers who prefer the "X-based Desktop Linux" mess?

This also echoes the rise of Linux for that matter in the 90s with technical users.

Sadly 90th have come and gone. Now it's Windows for non-technical users and MacOS for the users who want actually usable *nix.

Wayland - Beyond X (The H)

Posted Feb 15, 2012 23:29 UTC (Wed) by tstover (guest, #56283) [Link] (15 responses)

Non-technical user does not mean "mindless". A brain surgeon and a rocket scientist may simply want to browse the web when they get home from work, and potentially not also being heavily involved in "I.T." would therefore still be described as a non-technical user.

Guess that was a little too baited and off topic. It's just that I had to sit through another ipad consumerism drool session earlier today, and for some "crazy" reason thought there still might be someone left on LWN that would at least understand.

I'll concede that mac laptops are everywhere. Yes I see them outnumbering Linux ones at events (which I do go to), and yes the iron curtain of osx is now the apparatus of conformity and fashion that windows once was.

My comments were on how this came to be, not so much why it perpetuates.

I will say though, that you should at least open yourself to the possibility that maybe, just maybe - PHP, game, and google developers do not constitute the core of "linux users". Or for that matter that most people would even think of osx as a *nix. (although it is).

Regarding the mythical users, I'll put it like this. Everyday some kid somewhere discovering computing for the first time is hooked. That first program written, or that first system built, or that first network turned on, or that first cloud image deployed, or what ever it is they do now - is all it took. For the rest of their lives, that is all they will ever want to do. No one will ever dissuade them, lock them out, or shut them down. They can not walk away even if they wanted to. Whether or not F/OSS was involved in this pivotal moment, eventually it to enters into their bloodstream, and that to becomes a force far too powerful to walk away from. Given time, the thought of relinquishing the power of say Linux for a proprietary death sentence like osx, would be like a free man bowing down to a king.

How many of them are there? No one knows. We lurk in your schools and businesses. We hand out information and software to anyone who wants it. At air ports we see the sea of apple hardware, and actually desire not to be just like everyone else. At conferences we even show people how and why to overwrite the "X-based Desktop Linux mess" right on top of their windows and osx factory software, so that they too might also be free some day.

For what it's worth I've never said I was anti-wayland. My first post was stating that I think is going to help X development.

I'm really not even anti-osx. If we were in person, I would light up a cigar and open a bear with you, because if your actually reading my comments then you probably are either as bored or as distracted as I am.

Wayland - Beyond X (The H)

Posted Feb 16, 2012 9:14 UTC (Thu) by khim (subscriber, #9252) [Link] (14 responses)

I will say though, that you should at least open yourself to the possibility that maybe, just maybe - PHP, game, and google developers do not constitute the core of "linux users".

No? WTH NO? These are the people who write and run software on Linux five days per week. Very few of them develop anything for MacOS. Yet they overwhelmingly prefer to use Mac to develop said Linux software. Don't it say you something about current state of affairs? Note that few years ago they all had iPhones, today a lot of them use Androids - and not just Googlers (where it's obvious choice because most of them get Androids as a gift from Google), but people on the other side of the fence, too. This means that Apple does not have a monopoly on good design and Linux (the kernel) is not the culprit. But something else is for such things don't happen without a reason.

How many of them are there? No one knows. We lurk in your schools and businesses. We hand out information and software to anyone who wants it. At air ports we see the sea of apple hardware, and actually desire not to be just like everyone else. At conferences we even show people how and why to overwrite the "X-based Desktop Linux mess" right on top of their windows and osx factory software, so that they too might also be free some day.

Well, that's good explanation of where Linux developers come from, not where Linux users come from. And while developers are always valuable it's the lack of users which is problematic. Till you have certain amount of users hardware companies ignore you, they don't publish specs and don't explain how to use the goodies - thus you spend your limited developer's resources trying to reconcile changes in "latest and greatest" hardware (made without Linux in mind because there are so few users) with your software, it's often broken, etc.

Linux on server obviously is big enough to keep hardware companies interested while Linux on desktop is not. And as FreeBSD example shows server may not be to keep given OS alive and thriving.

Wayland - Beyond X (The H)

Posted Feb 16, 2012 14:52 UTC (Thu) by tstover (guest, #56283) [Link] (13 responses)

-Anyone developing on a laptop full time is due for some serious hand, neck, and back injuries. (this is not flame bate, but a very serious warning from experience of many)

-Android != Linux

-knowing a bunch a people who work at google that use mac laptops + using one yourself != no one uses X11 linux computers

-developing "for linux" on mac, either means one is running linux in a vm, remoted into linux, or is doing cross platform development. For instance I do cross platform development for *nix, windows, and yes osx. I still hate osx, and only put up with on the last stages of compiling and testing. I also hate windows, which is why I, like zillions of others use, - wait for it - linux computers.

-Sure apple and google are popular and having their days in the sun. Look beyond the main stream in all aspects of life, and you will find so much more. Seek individuality, beauty, nonconformity, nature, books, the human being, and you will find these things do not come from advertisements, top 40, and expensive fashion electronics.

-One thing I do like about wayland is trying to simplify the programming model. Reminds of of SDL with the concept of frame buffers in windows.

Wayland - Beyond X (The H)

Posted Feb 16, 2012 18:06 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link] (12 responses)

>-Anyone developing on a laptop full time is due for some serious hand, neck, and back injuries. (this is not flame bate, but a very serious warning from experience of many)

I've been doing it for more than 10 years. You can put a laptop on a table, you know.

Wayland - Beyond X (The H)

Posted Feb 16, 2012 19:24 UTC (Thu) by tstover (guest, #56283) [Link] (11 responses)

give it about another 6 years. Somewhere around year 17-18 is when I was forced to address the situation.

Sure just like "everything else" the con artists, lawyers etc, are out there with some just stupid comments, but that doesn't mean there isn't a real danger.

Of course this has much to do with how tall you are, how big your hands are, etc. Also most of the hand problems are actually do to unknowingly bending ones wrists while sleeping. Research has found people who type allot or play guitar for instance often start to do that. Eyes are another area to start taking care of.

Sure a quick session here and there is fine, but if your are going to sit down for the day its another story.

When I put one on a table, I put something under it to keep my neck from having to bend down, and use an external keyboard & mouse (preferably kinesis advantage and kensington trackball)

Leaning forward hunched over in the "oger position" will drop you for sure, given enough time. The blood flow in your neck being pinched off also exponentially increases your chances of a stroke.

These are not irrecoverable injuries though. It was actually the best thing that ever happened to me. After chiropractor + physical therapy, diet and exorcise put the lifestyle induced diabetes back in the closet. I've never been healthier.

No B.S. I tell these younger folks all the time about these things. If this is your career and you plan be able to do this for decades, then you really need to take care of your body. Stop eating the poisons, and get some real exorcise. Never underestimate how closely related your physical and mental health are either.

So in conclusion, X11 network transparency is a feature I like.

Wayland - Beyond X (The H)

Posted Feb 16, 2012 20:23 UTC (Thu) by khim (subscriber, #9252) [Link] (9 responses)

Leaning forward hunched over in the "oger position" will drop you for sure, given enough time.

This is kind of strange, because people who needed to read a lot used lecterns for centuries - and if you open your laptop case to about 130-140° you basically get the same thing on about optimal distance from eyes. It's not as if human body is designed to only ever see the sky, you know.

I think the problems with laptops starts when people try to uncritically apply rules designed for desktops (screen is vertical, top of the viewing screen at eye level, etc). But laptop's screen is way below you! If you'll constantly bend your body in the position to put your eyes at the level with the top of the screen I'll be surprised if you'll manage even few years! Good keyboard is harder to find but this is where YMMV significantly: some people have trouble even with good ergonomic keyboards, some can live with laptops just fine. I, for one, find ThinkPad's keyboard easier to deal with then traditional Keyboard+Mouse combo - because I don't need to move hands to use mouse: trackpoint is right in the middle of keyboard, after all (if I need precision then mouse is obviously better, I'm talking about web surfing, text editing, etc).

Wayland - Beyond X (The H)

Posted Feb 16, 2012 20:31 UTC (Thu) by dlang (guest, #313) [Link] (4 responses)

you can get full size keyboards with a trackpoint (or with a touchpad below the spacebar, almost as good)

Wayland - Beyond X (The H)

Posted Feb 16, 2012 20:46 UTC (Thu) by khim (subscriber, #9252) [Link] (3 responses)

you can get full size keyboards with a trackpoint

There are very few of them and they are no better then ThinkPad's built-in one, so what's the point?

or with a touchpad below the spacebar, almost as good

Not even close. Either you need to use thumb to move mouse pointer (way less precise then when you use index finger) or you need to move hands (and if I'll do that then I'll prefer traditional mouse over touchpad becase, again, it's more precise).

Wayland - Beyond X (The H)

Posted Feb 16, 2012 21:23 UTC (Thu) by dlang (guest, #313) [Link] (2 responses)

> There are very few of them and they are no better then ThinkPad's built-in one, so what's the point?

the rest of the keyboard is much better

Wayland - Beyond X (The H)

Posted Feb 16, 2012 21:53 UTC (Thu) by khim (subscriber, #9252) [Link] (1 responses)

the rest of the keyboard is much better

How come? Most of the look like this - basically laptop keyboard without a laptop. Not exactly sure how separation of keyboard from laptop can suddenly make it "much better".

Wayland - Beyond X (The H)

Posted Feb 16, 2012 22:45 UTC (Thu) by dlang (guest, #313) [Link]

two things.

1. having it be separate lets you position it better.

2. I was actually assuming that you would get a 'real' keyboard. In the past I've seen full-size M series type 'clicky' keyboards that have the erasermouse pointer in them.

by the way, I somewhat question if a thumb on a trackpad is really that much less accurate than the erasermouse joysticks, but I guess that's a personal preference :-)

Wayland - Beyond X (The H)

Posted Feb 16, 2012 20:43 UTC (Thu) by tstover (guest, #56283) [Link] (3 responses)

Using a lectern one is either standing, which has totally different muscle usage in the back (and you wouldn't do it all day), or has it closer to eye level than a laptop on a table. Reading is still a problem for me. Usually I have to sit a table to prop up my elbows and hold the book higher up. I'm an extreme case though. Sometimes certain car wreck injuries create the same problems for people. hunched over computer use leads to condition known as "upper cross syndrome", which can sometime be seen in swimmers and certain weight lifting patters. It's awful.

Wayland - Beyond X (The H)

Posted Feb 16, 2012 20:52 UTC (Thu) by khim (subscriber, #9252) [Link] (1 responses)

Using a lectern one is either standing, which has totally different muscle usage in the back (and you wouldn't do it all day), or has it closer to eye level than a laptop on a table. Reading is still a problem for me.

This may be the key difference. I have 20/20 vision (rarity today, I know) thus text on latop in aforementioned position is just perfect for me.

Wayland - Beyond X (The H)

Posted Feb 16, 2012 21:08 UTC (Thu) by tstover (guest, #56283) [Link]

> This may be the key difference. I have 20/20 vision (rarity today, I know)

as do I. (up until a few years ago it was even better than that). For many the issue is very much leaning forward to get closer to the screen. One of the many reasons I recommend monitor arms for desktop systems.

Torso height is the other big factor. If you sit with proper posture without bending your neck and your eyes can look close to straight ahead - then great! Many people have to sharply look down to do this which is unnatural to say the least. This causes the had to bend down, which eventually caused the back to slouch.

If you are reading this and you find your self doing this, please take the chance to start new habits.

Wayland - Beyond X (The H)

Posted Feb 22, 2012 7:39 UTC (Wed) by ssmith32 (subscriber, #72404) [Link]

I have a few co-workers that have their desks set up so they can stand up and work. I get the impression it's not so bad once you get used to it. Sort of like sleeping on a hard floor. It seems weird for a while, but doing that for a year or so improved some lower back issues I had quite a bit. People can stand for hours idling their time away on electronics... haven't you seen the lines at Apple stores :D ?

Wayland - Beyond X (The H)

Posted Feb 23, 2012 18:57 UTC (Thu) by sdalley (subscriber, #18550) [Link]

tstover, thank you for your very sensible comments about posture.

Wayland - Beyond X (The H)

Posted Feb 16, 2012 21:02 UTC (Thu) by cdmiller (guest, #2813) [Link]

This statement:

"and MacOS for the users who want actually usable *nix."

appears to fall under:

"And how we've entered realm of pure fantasy and self-delusion."

Wayland - Beyond X (The H)

Posted Feb 15, 2012 23:22 UTC (Wed) by HelloWorld (guest, #56129) [Link]

> It's not 1990 anymore. The concept of X over the network is no longer as strange as you would like to pretend. The inclusion of a bad implementation of this idea in MacOS is ample demonstration of this. The network GUI concept is fairly commonplace in the corporate Windows world. If anything, people should be improving these features in Linux rather than trying to abandon them.
Nobody is trying to abandon anything. The X.org server will still run on Wayland, and if application or toolkit developers choose to abandon support for X11, you should blame them and not the Wayland developers. OTOH, Wayland is what finally gives us a chance to invent a new remote rendering protocol that, unlike X11, doesn't suck for modern applications.

Wayland - Beyond X (The H)

Posted Feb 19, 2012 5:44 UTC (Sun) by AngryChris (guest, #74783) [Link] (2 responses)

You lost me here:

" For example you can count number of Macs (which don't support anything besides VNC) and PC laptops (which at least theoretically can support network transparency) on Linux conferences."

cbell@athena:~$ ls -ld /Applications/Utilities/X11.app
drwxr-xr-x 3 root wheel 102 Jan 13 00:23 /Applications/Utilities/X11.app
cbell@athena:~$ uname -a
Darwin athena.local 11.3.0 Darwin Kernel Version 11.3.0: Thu Jan 12 18:47:41 PST 2012; root:xnu-1699.24.23~1/RELEASE_X86_64 x86_64
cbell@athena:~$

I use X11 on my Mac all the time. I never use VNC.

Wayland - Beyond X (The H)

Posted Feb 19, 2012 6:39 UTC (Sun) by raven667 (subscriber, #5198) [Link] (1 responses)

What you are describing is the same state as wayland currently. You can display remote MacOS apps via VNC and you can display and remote X11 apps but the native display protocol is not X11 and the X server is not where the graphics drivers live.

Wayland - Beyond X (The H)

Posted Feb 20, 2012 12:41 UTC (Mon) by dgm (subscriber, #49227) [Link]

Exactly. You will be able to run X11 on top of Wayland, the same way you can get it on OS X or even Windows.

The X11 protocol will still be used, but basically for where it really makes sense: applications running remotely on another host. Local applications will work better (less round trips, less display artifacts, less processes and contexts switches involved) on Wayland.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 11:46 UTC (Tue) by Cato (guest, #7643) [Link] (2 responses)

I have actually tried X11 from Windows PCs to Linux - it was quite painful to get working, and extremely slow even on a 100 Mbps LAN - I know that's an implementation issue as it worked fine in the old days on 10 Mbps LANs but it shows that remote X11 isn't very useful for Windows to Linux remote access.

VNC and NX have both worked well, though setting up an NX server looked too painful to contemplate.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 14:04 UTC (Tue) by tstover (guest, #56283) [Link]

I do that almost every day. It's been working wonderful for years. Putty + xmingw or cygwin on most hardware of the last 5 - 7 years on 100 or 1000mpb ethernet is very very close to a local application. Multiple remote computers, text editors, browsers, gimp sessions, many users, all day M-F. No I'm not making this up. In fact it's even mixed in with the similar 'doz feature from 2008r2.

Sure this sucked back with nt3.5 + exceed + 10mbp or what ever...

and on my home LAN with linux on both sides it's also great. although I use local browsers and media players.

Wayland - Beyond X (The H)

Posted Feb 15, 2012 7:49 UTC (Wed) by ronnyadsetts (subscriber, #47268) [Link]

> I have actually tried X11 from Windows PCs to Linux - it was quite painful
> to get working, and extremely slow even on a 100 Mbps LAN - I know that's
> an implementation issue as it worked fine in the old days on 10 Mbps LANs
> but it shows that remote X11 isn't very useful for Windows to Linux
> remote access.

Running X remotely with ADSL connections both ends and speed is more than acceptable. That's ~1.8Mb/s upstream at one end and 800Kb/s upstream at the other end.

It's certainly much better to use than VNC over the same connection, both speed wise and graphical quality wise.

I suggest that something is wrong with your set up if the speed is unacceptable over 100Mb/s.

Losing network transparency for X to me would be losing one of it's biggest strengths.

Ronny

Wayland - Beyond X (The H)

Posted Feb 14, 2012 15:08 UTC (Tue) by daglwn (guest, #65432) [Link] (16 responses)

> You may show of the X over network as much as you want but the simple
> fact: people don't use it.

I use it. There are use cases that require it, including every machine accessed through a remote terminal.

Why do Wayland supporters keep insisting no one cares about remoting when lots of people say otherwise?

Wayland - Beyond X (The H)

Posted Feb 14, 2012 16:55 UTC (Tue) by drag (guest, #31333) [Link] (15 responses)

> Why do Wayland supporters keep insisting no one cares about remoting when lots of people say otherwise?

Why do Wayland bashers continue to ignore the obvious fact that Wayland will be perfectly capable of support X11 Networking?

There is the fear, of course, that developers may stop using X11 and use Wayland native thus you lose the ability to do X11.

The counter argument is easy:

Of course, if developers and users see value in X11 they will continue to use it and insist for it's support. So as long as people have value in it it will exist.

It is worth noting that your choices are not limited to just X11 or VNC...

Wayland - Beyond X (The H)

Posted Feb 14, 2012 17:20 UTC (Tue) by daglwn (guest, #65432) [Link] (14 responses)

> Why do Wayland bashers continue to ignore the obvious fact that Wayland
> will be perfectly capable of support X11 Networking?

No, it won't. It will be able to run an X server as a client. That's not the same thing.

> There is the fear, of course, that developers may stop using X11 and use
> Wayland native thus you lose the ability to do X11.

Bingo.

> Of course, if developers and users see value in X11 they will continue
> to use it and insist for it's support.

The whole point of Wayland is to get rid of X, isn't it? If not, then what is the point? I'm not opposed to dumping X, but the replacement has to be fully functional and that includes network transparency.

> It is worth noting that your choices are not limited to just X11 or
> VNC...

Every alternative I've tried to remote X performs worse. I don't think my situation is unique.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 17:59 UTC (Tue) by raven667 (subscriber, #5198) [Link] (5 responses)

> The whole point of Wayland is to get rid of X, isn't it? If not, then what is the point?

If I read the Wayland docs correctly the point is to simplify local graphics handling by providing direct access to the toolkits, X11 will be a major client application to this graphics stack for the foreseeable future. This is not about removing the X11 protocol its about not requiring the X protocol server to also drive the graphics display, separating X protocol handling from driving graphics cards.

> Every alternative I've tried to remote X performs worse. I don't think my situation is unique.

That seems odd because raw X11, even tunneled over a compressed SSH session, is the worst performer of the available protocols (except maybe for uncompressed VNC). RDP and ICA to a Windows system is significantly better, NX (and apparently xpra which I haven't personally used) makes a big difference for X and there are other protocols such as SPICE and the VMware one which are also better, in pretty much every way. Many of these protocols can even do individual window redirection and not just full desktop remoting.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 18:48 UTC (Tue) by daglwn (guest, #65432) [Link]

All of those alternatives also require running some kind of networking process on the remote machine. I don't have to do that to use ssh -X. I don't have admin privileges on most application servers and the sysadmins tend to get testy if I install local servers.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 23:15 UTC (Tue) by nix (subscriber, #2304) [Link] (3 responses)

Why do people keep saying RDP is any good? It's dire. I've used it over both limited-bandwidth and high-latency networks, and the number of whole-framebuffer repaints it does when only a few characters have changed is beyond a joke. It's usable -- just. I don't really think 'usable -- just' is a good target to aim for.

Wayland - Beyond X (The H)

Posted Feb 15, 2012 9:21 UTC (Wed) by Fowl (subscriber, #65667) [Link] (1 responses)

I'd be interested in your experience.

The only time I've seen full repaints are with less aware toolkits, eg. Swing. Chrome's content area frame is opaque to the remoting stack too and is treated as a bitmap because of the way that they've done their buffer management/sandboxing.

(Assuming both the client and server were running NT 5+ clients - if there's enough of a feature missmatch then RDP will fallback to a vnc-like scheme, which isn't very good because it's not often used.)

Also, Citrix and other 3rd parties have made enhancements that some people may not distinguish from vanilla rdp.

Wayland - Beyond X (The H)

Posted Feb 15, 2012 15:36 UTC (Wed) by nix (subscriber, #2304) [Link]

I've seen it on Windows->Windows connections with Swing, VNC (sigh), XWin, and even MS Office 2008-or-whatever-number-it-is at the other end. Not every repaint is a full repaint: they seem to happen almost at random.

Wayland - Beyond X (The H)

Posted Feb 15, 2012 19:02 UTC (Wed) by jedidiah (guest, #20319) [Link]

RDP is not bad when you consider the alternatives.

It certainly beats anything that's being proposed by anyone pushing Wayland.

Short of X itself, all of the alternatives to RDP are even more dire.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 19:41 UTC (Tue) by drag (guest, #31333) [Link] (7 responses)

> No, it won't. It will be able to run an X server as a client. That's not the same thing.

It will be able to connect and display X clients. That's X11 networking.
Why is it such a surprise that it's going to run a program to provide X functionality?

It's like saying you can't browse the web with Wayland because you need to use a web browser.

> The whole point of Wayland is to get rid of X, isn't it?

No.

The whole point of Wayland is to provide a modern and relatively easy to platform for writing graphical programs. It will provide the ability for application developers to use direct rendering for their applications using whatever method or API is best suited for the application or the audience.

This is something that is not possible with XFree DDX we use today and X11 is one of those methods that application developers can choose to use.

It only intends to get rid of the need of XServer running you hardware. You can still have your XServer.

Saying that Wayland is intended to get rid of X11 is like saying that Wayland intends to do away with HTTP.

> If not, then what is the point? I'm not opposed to dumping X, but the replacement has to be fully functional and that includes network transparency.

Ok?

Trying to putting it into perspective:

I have 'network transparency' with Windows 7. I don't know why people that is X11 is so special or so required if you want to use remote applications.

It's not the only game in town and it's been a VERY VERY long time since it was even a good one. The world has moved on. The reason is that we still have relatively high latency links. We have lots of bandwidth that is plenty for running remote applications, however latency have not dropped by the same amount. Latency is what kicks X11's ass and is something that X11 is very poor at dealing with. All of which means that X11 is not suitable solution for what most people use remote apps for... which is things like implementing VDI and accessing applications over the internet over high latency links.

Keep in mind a _A_LOT_ of people use remote apps. My dad uses remote apps. 90% of people at my work use remote apps. Teachers, students, accountants.. all sorts of people use networking for remote GUI apps. They just don't use Linux, nor do they use X11 to do it.

Maybe X12 will make it useful once again. But keep in mind that it's going to be a lot easier to implement X12 support on top of Wayland then it's going to be with Xfree. This is because you can easily keep your backward compatibility with a X11 server integrated in Wayland while vastly reducing the overhead needed to implement a X12 server...

Wayland - Beyond X (The H)

Posted Feb 14, 2012 19:45 UTC (Tue) by dlang (guest, #313) [Link] (1 responses)

>> No, it won't. It will be able to run an X server as a client. That's not the same thing.

> It will be able to connect and display X clients. That's X11 networking.
Why is it such a surprise that it's going to run a program to provide X functionality?

make it possible to run a Wayland app and display it on an X display and you have what people need.

Then we can talk about the efficiency and performance of the Wayland remote capability. If all it does is shove full-resolution screenshots to the X display, it's not going to work very well.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 23:19 UTC (Tue) by nix (subscriber, #2304) [Link]

make it possible to run a Wayland app and display it on an X display
That's not that important. What matters is two things, really:

- That an X11 client for Wayland exists such that a Wayland server can display remote (and local) X apps. This seems overwhelmingly likely to be written or the transition process would be impossible. (It also seems likely that such a client would be transparently-started and per-window, making X apps look indistinguishable from native Wayland apps).

- That something exists such that remote Wayland clients can be displayed on local Wayland servers with reasonable efficiency (comparable to or better than current X, not giant bitmap-hurling horrors like VNC). So far I have received a mass of conflicting opinions regarding whether this is desirable or even possible. If this is implemented as a bunch of toolkit-level hacks, then I very much hope that no applications ever come to support Wayland except via those toolkits, because Wayland apps that only work locally will break the illusion that all machines are one machine that I very much value. (X is surprisingly good at this. I've watched full-motion 1680x1050 video over X connections on gigabit ethernet and it's worked so well that I didn't notice that I was running remotely until the video ended. It falls over in a heap if SSH tunnels are involved, of course -- SSH just can't transfer that much information that fast, nor was it ever designed to.)

Wayland - Beyond X (The H)

Posted Feb 14, 2012 20:06 UTC (Tue) by khim (subscriber, #9252) [Link] (4 responses)

It's even worse:
The reason is that we still have relatively high latency links. We have lots of bandwidth that is plenty for running remote applications, however latency have not dropped by the same amount.

Yup. Exactly. What's worse: in the future situation will go from bad to worse, the to the worst.

Imagine how the state of the art XXV century technology will look like. Bandwidth will probably be astronomical. Enough to shove 7680×4320 frames 120 times per second and more. Latency... Worst-time latency will still be 67ms on average, 133ms worst case. Calculations are trivial: 20000km-40000km round-trip (average 20000km, worst case 40000km), 300000km/sec, you do the math.

Why people still are trying to invent stupid schemes to reduce bandwidth requirements by adding more round-trips is beyond me: these schemes will be useless pretty soon, why try to add them to the new protocol?

Wayland - Beyond X (The H)

Posted Feb 14, 2012 23:05 UTC (Tue) by Cyberax (✭ supporter ✭, #52523) [Link]

Well, unless we get FTL :) Those guys with neutrinos are clearly up to something.

But yeah, light-speed lag is going to the limiting factor for a long time.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 23:26 UTC (Tue) by nix (subscriber, #2304) [Link] (1 responses)

Worst-time latency will still be 67ms on average, 133ms worst case
Hah. I wish. I use X (a bit, it is too slow to be useful most of the time) from my parents' house, in the wilderness, a whole ten miles from a major tourist spot in Yorkshire, so obviously the back of beyond (that's millions of centimetres, nobody could hope to string a cable that far). Satellite link to geosynch and back, then down to Italy, then back to England. 660ms RTT. This is the only broadband they are likely to get for at least the next twenty years. And they are relatively lucky: at least they didn't have a hill stopping them from seeing geosynch. What people in the Scottish highlands are doing for connectivity I have no idea, but it is probably not pleasant.

I don't expect any remote display protocol to cope with that sort of latency.

Why people still are trying to invent stupid schemes to reduce bandwidth requirements by adding more round-trips is beyond me: these schemes will be useless pretty soon, why try to add them to the new protocol?
Who on eath is trying to do that? Reducing round-trips has been core to X protocol optimization since the start: anyone trying to increase them (excepting only perhaps do-only-once stuff which is amortized to zero) is a nutter. However, not even local network bandwidth is yet free, and with 10GbE equipment at the cost it is now ($4000 for a basic switch!) I'd expect a long long wait before that becomes common enough that we can consider that it is even ten times cheaper than it is now.

Wayland - Beyond X (The H)

Posted Feb 15, 2012 0:42 UTC (Wed) by Cyberax (✭ supporter ✭, #52523) [Link]

>I don't expect any remote display protocol to cope with that sort of latency.

But RDP can do this, it'll be uncomfortable but usable. I had to use RDP a couple of times on congested GPRS/EDGE links with similar latencies.

Wayland - Beyond X (The H)

Posted Feb 15, 2012 19:02 UTC (Wed) by alankila (guest, #47141) [Link]

We should seriously study if we could just radiate information on tight beam straight through the Earth, presumably through some kind of fixed networking stations. Might get that worst-case latency down to 12000 km light time (40 ms). At least some particles are going to pass through Earth; difficulty may lie in both generating them headed to the right direction, and with detecting them...

Wayland - Beyond X (The H)

Posted Feb 14, 2012 4:12 UTC (Tue) by Cyberax (✭ supporter ✭, #52523) [Link] (58 responses)

It's already possible to run a rootless X inside Wayland.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 4:26 UTC (Tue) by dlang (guest, #313) [Link] (57 responses)

Yes, but if programmers shift to using a Wayland based library instead of an X based library it won't do you any good to be able to run X inside Wayland.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 6:34 UTC (Tue) by raven667 (subscriber, #5198) [Link] (56 responses)

Is an assertion that is not backed up by any facts that I am aware of. Do you think the major toolkits will rip out X11 support the day Weyland hits 1.0 in some sort of mass suicide pact? x11 will still be supported for cross platform applications for the foreseeable future. It takes a great galloping leap of logic to assume otherwise.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 7:44 UTC (Tue) by dlang (guest, #313) [Link] (47 responses)

I wouldn't expect a major, cross platform library to ditch X11 support

but unless there's a library that works with Wayland and is perceived to be better than X11, what's the point of creating Wayland?

and I would not expect a library created for Wayland to include support to run on X11 as well. first off it would add significant complexity to the library (probably undermining the performance on Wayland), but also any support that does exist is probably not going to be well tested, after all the people selecting that library are selecting it because they want to use Wayland.

In addition, there are a lot of people claiming very publicly that cross platform compatibility is holding back development of Linux (I don't just mean LP with systemd and journald, or GNOME talking about needing to be their own OS, not just a layer on top of Linux, but these are some of the biggest examples)

So why would I expect that cross platform compatibility is going to be so high on everyone's priority list?

Statements about how features haven't been used by anyone for a decade when people are using them daily don't reassure me either, for some strange reason ;-)

and this completely ignores any possibility of a ditry trying to create some tools that would be hard for other distros to use without jumping ship and making Wayland their display manager as well.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 10:23 UTC (Tue) by farnz (subscriber, #17727) [Link] (46 responses)

The point of Wayland is that even if all the Wayland clients are rootless X servers, merging input handling into the compositor is a net gain. It means that whatever transformation is applied to the output can also be applied to the input seamlessly.

And maybe that's where we'll end up; native Wayland is used for the compositor, and all the applications are X11 based, running on rootless X servers.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 17:16 UTC (Tue) by rqosa (subscriber, #24136) [Link] (45 responses)

> merging input handling into the compositor is a net gain

But X11 doesn't inherently require a separate window manager / compostiting manager. For example, MacOS X's rootless X server (and similar ones, such as Exceed for Windows) will use the native window system's window borders if there is no X window manager running. So, there's no reason in principle why there can't be an X server for Linux that has its own built-in compositing window manager.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 17:49 UTC (Tue) by raven667 (subscriber, #5198) [Link] (44 responses)

I imagine that this idea was considered and was found to be too horrible to entertain.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 18:51 UTC (Tue) by rqosa (subscriber, #24136) [Link] (43 responses)

Care to explain what's wrong with it? The way I see it, in the local-applications-only use case, an X server with built-in compositing together with a widget toolkit capable of direct rendering (DRI2 and OpenGL ES and/or OpenVG) would give all the features that Wayland is supposed to have, while being a less disruptive change.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 19:17 UTC (Tue) by raven667 (subscriber, #5198) [Link] (26 responses)

Well for one that would be a major architectural change for the x.org server that might not be appreciated by all the consumers of it (and certainly not by competitive compositors/window managers). Which compositing window manager would you make the x.org server standard? That's the horrible part. It also is more complicated, if the apps are going to direct render with DRI2, then what benefit is there to running everything through the X server rather than treating the X server as just another direct render client. The Wayland approach is architecturally more simple and robust.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 20:37 UTC (Tue) by rqosa (subscriber, #24136) [Link] (25 responses)

> Well for one that would be a major architectural change for the x.org server that might not be appreciated by all the consumers of it (and certainly not by competitive compositors/window managers).

But Wayland is also a major architectural change that requires existing compositing window managers to adapt to it. (If I understand correctly, it requires existing compositing window managers to be rewritten as servers that talk the Wayland protocol.)

> Which compositing window manager would you make the x.org server standard? That's the horrible part.

Then Wayland is equally "horrible" (because it combines a compositing window manager and a display server into a single application, just like what I'm suggesting).

if the apps are going to direct render with DRI2, then what benefit is there to running everything through the X server

"Everything" doesn't run through the X server — that's what "direct rendering" means. But we still need a server for input handling and window management. And the extension system in the X protocol allows for (local-only) direct-rendering clients (DRI2 + EGL or GLX) to coexist with (remote or local) indirect-rendering clients (XRender or AIGLX), with possibly even the exact same applications able to run in either direct or indirect mode. (For example, an OpenGL application can run on either GLX+DRI2 or AIGLX. Maybe an accelerated-indirect version of EGL would give good enough performance for running OpenVG applications remotely?)

> The Wayland approach is architecturally more simple and robust.

How is it more simple? If you're "treating the X server as just another direct render client", then you still need to have an X server present, and you also need for input events to pass through two servers (rather than one) before reaching the client. And if you're using only direct-rendering clients, you could have a minimal X server that omits extensions you're not using (e.g. XRender). (And even if you don't omit them, it seems like the only cost would be some more memory pages allocated to the X server, which would then get paged out because they're not actually being used.)

Wayland - Beyond X (The H)

Posted Feb 14, 2012 20:48 UTC (Tue) by raven667 (subscriber, #5198) [Link] (2 responses)

Wayland is a new component and isn't going to require architectural changes to the x.org server to support. Compositing window managers will need to be ported to Wayland, and one is being developed alongside for reference but that one isn't mandated or baked in so there can still be competitive window managers. I think that is where the confusion is, if I understand the architecture correctly, the compositing window manager is still a replaceable component.

Wayland is being designed and implemented by the same people who designed and implemented DRI2, AIGLX and X11 itself so I think they probably know more than I do about what is the most sane architecture to implement. As you describe, if you have a minimal X server that is only handling direct rendering clients, then you have Wayland. Call it X12 in your head if you like.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 21:31 UTC (Tue) by rqosa (subscriber, #24136) [Link] (1 responses)

> if I understand the architecture correctly, the compositing window manager is still a replaceable component.

If by "replaceable component" you mean a separate process, then no, it isn't. See this blog posting by the developer of Compiz:

Unlike X11 where we have a bunch of clients push to a display server, and then an external window manager / compositing manager process, in wayland we have a single process for a display server and compositing manager. There is no window manager, all of that is pushed to the clients (move, resize, etc).
My understanding is that there is a library ("libwayland") which implements the server side of the Wayland protocol, and compositing managers (such as Compiz, KWin, etc.) are supposed to be rewritten as daemons that talk the Wayland protocol to clients (with the actual server implementation being in libwayland, which they link to — I'm guessing that the program is supposed to register several event-handler callbacks with libwayland, then call into a main-loop function that libwayland provides, and the provided callbacks will be called in response to things like input events and start/stop of client programs).

It just seems to me that, in principle, it should be possible to adapt the X.Org Server so that it becomes just like this — a shared library to be embedded by compositing manager programs (or alternately, have the the X server remain a daemon but add a plugin API to it that allows for compositing managers to be developed as in-process plugins to the X server).

Wayland - Beyond X (The H)

Posted Feb 14, 2012 22:46 UTC (Tue) by farnz (subscriber, #17727) [Link]

In principle, you could indeed rewrite the existing X server to provide a suitable interface to plug a compositor in, much like Wayland. But if you look at the X.org codebase, you realise that it's a complete mess, despite all the cleanup work that's been done, and that this is a harder task than the way round that Wayland has done it (where X becomes just another client of a Wayland system).

The people doing this aren't complete idiots - they're the people who are keeping X11 going today. They're hitting the pain points of the current codebase, and Wayland began as an experiment to see if they could do something useful with the parts they already have, without a long silent period where they're simply refactoring all the cruft out of the existing X server codebase. Turns out, the answer's yes - it could just as easily have been "no, not interesting unless you refactor X.org".

Wayland - Beyond X (The H)

Posted Feb 14, 2012 20:52 UTC (Tue) by wmf (guest, #33791) [Link] (2 responses)

> But Wayland is also a major architectural change that requires existing compositing window managers to adapt to it. (If I understand correctly, it requires existing compositing window managers to be rewritten as servers that talk the Wayland protocol.)

Actually, I think people expect existing compositors to die and everyone will just use Weston. Although we may see Weston-Unity and Weston-GNOME forks.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 21:06 UTC (Tue) by raven667 (subscriber, #5198) [Link]

Yeah, that's what I would expect, GNOME and KDE to have their own forks or to implement Wayland directly in kwin and Mutter, similar to the evolution of compiz.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 22:30 UTC (Tue) by tjc (guest, #137) [Link]

> Although we may see Weston-Unity and Weston-GNOME forks.

And certainly Weston-Cinnamon. :)

Wayland - Beyond X (The H)

Posted Feb 15, 2012 0:17 UTC (Wed) by Cyberax (✭ supporter ✭, #52523) [Link] (18 responses)

Have you _seen_ the X codebase?

It's ancient. It has all kinds of arcane features (like an x86 real mode emulator for legacy BIOS drivers!!!). Moving compositor inside Xorg would require huge refactoring of a giant codebase which only a handful of people on Earth really understand completely.

Wayland on the other hand is small and concise. You can easily write a Wayland protocol implementation in a _weekend_ (spoiler: I have a pure Java implementation).

Adapting existing frameworks for Wayland is also not terribly hard. QT, GTK and EFL already have alpha-quality implementations which covers the majority of currently used frameworks.

Wayland - Beyond X (The H)

Posted Feb 15, 2012 0:56 UTC (Wed) by rqosa (subscriber, #24136) [Link] (5 responses)

> It has all kinds of arcane features (like an x86 real mode emulator for legacy BIOS drivers!!!).

Who cares? Any system that's capable of running Wayland (which requires Linux and KMS) doesn't use that part of X.Org anyway.

> QT, GTK and EFL already have alpha-quality implementations which covers the majority of currently used frameworks.

That's the problem, right there. We wouldn't be complaining if Wayland were just a piece of infrastructure for the X server to use — but if the toolkits are going to migrate to the native Wayland protocol, then we lose the ability to use all of our current remoting solutions (NX and xpra) other than VNC (or potentially some future slightly-improved variation of VNC, which still will have bad performance compared to things like RDP).

Wayland - Beyond X (The H)

Posted Feb 15, 2012 1:00 UTC (Wed) by rqosa (subscriber, #24136) [Link]

s/running Wayland/running the existant Wayland compositing servers/

Wayland - Beyond X (The H)

Posted Feb 15, 2012 15:52 UTC (Wed) by nix (subscriber, #2304) [Link] (3 responses)

Quite so. Awesome local performance is great -- I don't want to argue against it. But gaining awesome local performance (which, for most apps I use, is moving things from too-fast-to-see to too-fast-to-see) at the cost of *all* remote performance (moving things from slight-delay to oh-now-I-can't-do-my-work-at-all) is a very definite loss.

Wayland - Beyond X (The H)

Posted Feb 16, 2012 10:42 UTC (Thu) by intgr (subscriber, #39733) [Link] (2 responses)

> moving things from too-fast-to-see to too-fast-to-see

Well, not in my experience. I can sense a noticeable delay in most GUI applications. Click on a menu and it takes some time for the dropdown to appear. Click on a button and it takes noticeable time for the next window to pop up.

I'm consistently amazed when I see my coworkers using their well-maintained Windows XP machines. It's as if menus pop up before they're actually clicked. Simpler applications start up with no visible delay.

Maybe some of my problem is the fact that I use weak-ish graphics cards in a dual monitor configuration. Maybe it's suboptimal drivers. Maybe it's inefficient GUI toolkits. Maybe it's power management. But sure as hell, my X11 desktop has crappy interactivity.

Wayland - Beyond X (The H)

Posted Feb 16, 2012 23:13 UTC (Thu) by rqosa (subscriber, #24136) [Link] (1 responses)

Windows XP might feel more responsive because it doesn't use compositing, which can have a performance cost (especially on older GPUs). On my 4.5-years-old hardware, programs feel slightly more responsive when KWin's compositing is turned off than when it's on.

(And of course there's also the possibility that the slowness you're experiencing is the fault of the applications and/or the widget toolkits, rather than the underlying window system. For example, I've noticed that Qt4 seems a lot slower than Qt3 when running on older CPUs.)

Wayland - Beyond X (The H)

Posted Mar 1, 2012 14:01 UTC (Thu) by nix (subscriber, #2304) [Link]

Quite. I'm still using KDE3 (well, Trinity). Too-fast-to-see. I can't see what compositing brings to the table other than wobbly windows, which I really don't care about.

Wayland - Beyond X (The H)

Posted Feb 15, 2012 15:51 UTC (Wed) by nix (subscriber, #2304) [Link] (11 responses)

The x86 real mode emulator is there to handle userspace modesetting for cards for which initialization requires the execution of x86 BIOS code without requiring the machine to be an x86 (or to allow it to happen while the x86 is in long mode), and which don't have enough information provided in another form (e.g. as tables) to permit reliable initialization in any other way. Among other things this includes all pre-ATOMBIOS Radeon cards, so this piece of ugly is not going away until userspace modesetting does.

If Wayland is going to drop that, it's dropping UMS as well (which is quite plausible).

Wayland - Beyond X (The H)

Posted Feb 15, 2012 16:02 UTC (Wed) by Cyberax (✭ supporter ✭, #52523) [Link] (10 responses)

Wayland have never used UMS (and I don't think it ever would), it relies on KMS.

Wayland - Beyond X (The H)

Posted Feb 15, 2012 18:37 UTC (Wed) by dlang (guest, #313) [Link] (9 responses)

well, if you were to take the X codebase and drop every driver that doesn't use KMS, and then further limit your target platform to Linux (the only OS with KMS), there is a HUGE amount of code that you could drop.

but as long as you want to provide _any_ output on these other systems, you need to have all this 'cruft' in the codebase.

Wayland - Beyond X (The H)

Posted Feb 15, 2012 21:10 UTC (Wed) by khim (subscriber, #9252) [Link] (8 responses)

but as long as you want to provide _any_ output on these other systems, you need to have all this 'cruft' in the codebase.

Well, the logical conclusion says that there will be no Wayland support for these systems. Not a big loss.

P.S. You can as well complain that Wayland will not support X terminals.

Wayland - Beyond X (The H)

Posted Feb 16, 2012 6:34 UTC (Thu) by jmorris42 (guest, #2203) [Link] (7 responses)

> Not a big loss.

Uh huh. So no cross platform, no network transparency and mixing mechanism and policy. Fail, fail and fail. Not seeing any Win to balance it so Meh. Sure Lennart Pottering isn't involved in this project? Seems to have the same trademark militant rejection of every single philosophical underpining that made UNIX/X/GNU a winner.

And notice how all three are under active attack, Pottering is hammering away at the core assumptions of UNIX, Wayland is directly aimed at the key product differentiators that made X a legend and we are suddenly seeing concentrated attacks on GNU and the GPL. Hmmm.. If you can't beat em straight up, make a long march through thier institutions and gut em from within? I guess the endpoint pf all this is a GNOME 4 written in Mono, locked to local display on Wayland and licensed under BSD. Bah.

Wayland - Beyond X (The H)

Posted Feb 16, 2012 9:38 UTC (Thu) by khim (subscriber, #9252) [Link] (3 responses)

Sure Lennart Pottering isn't involved in this project?

Nope. Net yet, anyway.

Seems to have the same trademark militant rejection of every single philosophical underpining that made UNIX/X/GNU a winner.

s/winner/looser/

UNIX/X/GNU already lost the workstations, desktops, tablet and mobile. What's left? Ah, server… Care to explain where exactly X11 is a big win on server? AFAICS server Linux thrives in spite of X11, not because of X11.

I guess the endpoint pf all this is a GNOME 4 written in Mono, locked to local display on Wayland and licensed under BSD.

Well, if this will finally make the platform usable for normal people then it still will be a win.

So no cross platform, no network transparency and mixing mechanism and policy. Fail, fail and fail. Not seeing any Win to balance it so Meh.

The win is supposedly simpler and more robust codebase which will make it possible to write snappier, more fluid interfaces. If the promise will be fulfilled then the whole exercise will be justified, otherwise Wayland will be just an article in a Wikipedia.

Wayland - Beyond X (The H)

Posted Feb 16, 2012 12:12 UTC (Thu) by rqosa (subscriber, #24136) [Link] (2 responses)

> s/winner/looser/

It's not a "loser" for the many people who are using it right now.

> Well, if this will finally make the platform usable for normal people then it still will be a win.

If making it usable for some hypothetical "normal people" means making it unusable for the current users, that's not a win.

Wayland - Beyond X (The H)

Posted Feb 16, 2012 14:49 UTC (Thu) by khim (subscriber, #9252) [Link] (1 responses)

If making it usable for some hypothetical "normal people" means making it unusable for the current users, that's not a win.

It's free software. Current users can use current version of the stack or even create a fork - it's to them.

If Wayland will lose few existing users and gain a lot new users it will still be a net win. Yes, it's a gamble, but it's well-considered gamble.

Wayland - Beyond X (The H)

Posted Feb 17, 2012 9:38 UTC (Fri) by alankila (guest, #47141) [Link]

To co-opt certain language that's been popular recently, Linux currently is for the 1 %, not the 99 %. You can afford to lose some of that 1 % if you gain nearly any fraction of the 99 %: number of users would increase dramatically.

And as you correctly observed earlier, traditional Linux's biggest problem right now is that it does not have enough users, and that is an existential threat because it means no games, no hardware support, and eventually not even computers to run it on. We probably need customer demand for Linux to even have a long-term future. This problem must be fixed, and we're going to need kick-ass local graphics among other things. I'm sure that we can do better than X network transparency also, because it's at least possible when we finally start to seriously look into doing it at toolkit level.

Wayland - Beyond X (The H)

Posted Feb 16, 2012 14:01 UTC (Thu) by michich (guest, #17902) [Link]

> So no cross platform, no network transparency and mixing mechanism and policy. [...] Sure Lennart Pottering isn't involved in this project?

Do you realize that the guy gave you workable network transparency for audio?

Wayland - Beyond X (The H)

Posted Feb 16, 2012 18:02 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link] (1 responses)

>I guess the endpoint pf all this is a GNOME 4 written in Mono, locked to local display on Wayland and licensed under BSD. Bah.

You are describing Android. And yeah, it's a win - it's the most popular OS for end-users right now.

Wayland - Beyond X (The H)

Posted Feb 16, 2012 22:47 UTC (Thu) by rqosa (subscriber, #24136) [Link]

But those users already have an operating system they like (one that is already FLOSS, even). So how can there be there any good reason to change GNOME (and the underlying infrastructure it uses) to make it more like Android?

(And if your answer is something like "Android isn't really FLOSS", then the real solution to that is that we need to have unlocked hardware capable of running the community-developed Android rebuilds like CyanogenMod and Replicant on, and to educate the public about why locked-down hardware is bad.)

Wayland - Beyond X (The H)

Posted Feb 14, 2012 22:12 UTC (Tue) by farnz (subscriber, #17727) [Link] (15 responses)

That's basically what X on Wayland is - another way to look at Wayland is that it's the construction kit to build such an X server, and as we're putting all the bits together, we might as well expose the lower layers so that you can bypass X if you want to.

Wayland's protocol is a spiritual subset of X11 - it's basically clipboard data exchange, input and DRI2, without the rest of the X11 bits. The trouble with building this atop X11 is that the interactions between X11 core protocol and the subset that Wayland provides are nasty (for example, X11 has traditional input events that "know" that you only ever have one mouse position and a set of buttons - what is the "correct" mouse position if you're doing multitouch input? Where did you touch when you make the "zoom in" gesture - the start point? The end point? - noting that "neither" and "both" aren't acceptable answers to core X input), and you have to get them right to truly claim to be "an X server with built-in compositing". The rules around subwindows are another example

Wayland takes it in a different direction - throw away the bits that are hard to implement yet rarely used, and make a Wayland client (the X server) implement them if they're wanted. If network transparency really is a critical feature, someone will step up to implement it - whether by doing the work themselves, or by paying someone such as Red Hat to include it.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 22:24 UTC (Tue) by dlang (guest, #313) [Link] (3 responses)

I understand that X supports multiple mice on a screen, and I've seen stuff recently about it supporting multitouch as well.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 23:11 UTC (Tue) by farnz (subscriber, #17727) [Link] (2 responses)

It has some deep seated issues around the way multiple mice and multitouch interact with "traditional" X clients - if we could just assume that all X applications will be rewritten to assume multitouch support, we could get rid of a decent collection of interesting special cases.

There are at least three cases that XI2 just can't get right when interacting with traditional X clients:

  1. When I have multiple simultaneous touch points, where is the "correct" location for the emulated single touch? There's no right answer for this - it could be one or more of the individual touch points, or it could be the position of the cursor, or it could be an average of the locations of the touch points. Exactly which one is right is a function of the underlying UI, which X can't see into.
  2. If the user does a "global" gesture, you need to tell the client "I know I've been feeding you motion and touch events, but actually, they've turned out to be a global gesture. You need to forget that they happened". The alternative is to not send the events at all until touch up - painful if you're trying to drag an icon.
  3. With multiple input devices, how do I synthesise the single mouse that a traditional X client expects? Whatever I do is going to fail in some corner cases - either there will be button events that don't correspond to your expectations (e.g. because I simply serialised two mice worth of events into one stream, and mouse one was trying to drag an icon when mouse 2 came in and double-clicked a different icon), or there will be a lack of events when you expected some (e.g. because mouse 1 is lurking over a window, but you've moved mouse 2 over it as well, and have forgotten about mouse 1).

Wayland - Beyond X (The H)

Posted Feb 14, 2012 23:38 UTC (Tue) by khc (guest, #45209) [Link] (1 responses)

Curious comment from a bystander, how does wayland plan to address these issues? Specifically 1) and 2).

Wayland - Beyond X (The H)

Posted Feb 15, 2012 12:15 UTC (Wed) by farnz (subscriber, #17727) [Link]

Wayland's way of addressing them is simple - it doesn't have a historic protocol to worry about, so it can assume that you are multitouch aware. It handles them the same way XI2 does when dealing with XI2 clients.

Specifically addressing the three points I made earlier:

  1. When you have multiple touch points, Wayland reports all the touch points individually, including their locations, and the client is expected to work out what that means. For example, it could highlight a button if any of the touch points are within that button, if that's appropriate for this client.
  2. Touch event streams are terminated with either a "touch complete - go process" or "actually, that was the beginning of a global gesture - ignore it" event. This lets applications provide feedback to the user as the touch happens (e.g. drag pointer instead of normal pointer), but delay taking action until the touch completes.
  3. You get events from each input device individually, so the client can distinguish event streams from multiple mice.

Wayland - Beyond X (The H)

Posted Feb 15, 2012 10:36 UTC (Wed) by renox (guest, #23785) [Link] (8 responses)

> Wayland's protocol is a spiritual subset of X11

The issue is at which point a subset is good enough?

> If network transparency really is a critical feature, someone will step up to implement it

Which network transparency?
Inefficient network transparency based on buffer would be easy to add sure, but efficient network transparency to Wayland would need to change a lot Wayland, adding the equivalent of XRender in the API (especially the part used for text rendering) so that toolkits would need to be significantly changed also (may not be too difficult actually for a multi-platform toolkit which can also use X11, but for a Wayland-only toolkit..).

Wayland - Beyond X (The H)

Posted Feb 15, 2012 12:25 UTC (Wed) by farnz (subscriber, #17727) [Link] (7 responses)

So, the expected path is that Wayland handles local compositing and input. There will probably be a VNC-alike inefficient remoting protocol for native Wayland at some point, assuming anyone gets interested (which is a safe assumption, as VNC exists for platforms like Windows that didn't have a native remoting protocol from day 1).

Something else, toolkit specific since that's the layer at which you have semantic information to do a good job, will handle efficient network transparency. You can already do this today for GTK+ by writing a new GDK backend that produces a network stream that can be fed back into GDK at the other end.

And note that (because Wayland's protocol is basically a subset of X11's) it's entirely possible already to produce an X11 toolkit that cannot be efficiently transported over the network. I've already encountered proprietary X11 applications whose custom toolkits handle X11 by sending repeated full-window bitmaps over the wire, after all, so it's not like X11 is a guarantee that you'll do something network-efficient.

Wayland - Beyond X (The H)

Posted Feb 15, 2012 13:54 UTC (Wed) by renox (guest, #23785) [Link] (5 responses)

> Something else, toolkit specific since that's the layer at which you have semantic information to do a good job

Toolkits have the information yes, but hopefully they can use a generic protocol (say Wayland 2.0 with something like Render added), not a toolkit specific protocol: it would be a compatiblity nightmare..

Wayland - Beyond X (The H)

Posted Feb 15, 2012 15:29 UTC (Wed) by farnz (subscriber, #17727) [Link] (4 responses)

The problem right now is that we don't actually know what that generic protocol should look like. XRender is better than core X11, but is still neither good enough for modern toolkits (which are moving to OpenGL to escape the problems with XRender), nor as efficient as something like NeWS was.

Hopefully (yes, I'm an optimist), we will see some experimentation with remoting protocols as a fallout from Wayland, and eventually find a better compromise than "accept the limitations of X11".

Wayland - Beyond X (The H)

Posted Feb 16, 2012 13:08 UTC (Thu) by renox (guest, #23785) [Link] (3 responses)

> XRender is better than core X11, but is still neither good enough for modern toolkits (which are moving to OpenGL to escape the problems with XRender)
Which issue are you talking about? The speed of the rendering?
I expect that any protocol which is network efficient (WAN) will reduce the rendering speed, it's a tradeoff..

> nor as efficient as something like NeWS was.
Sure, but I don't think that it's possible to keep current toolkits and have them implement something like NeWS or Fresco/Berlin.

But current toolkits already support X11 so if there was a Wayland2.0 which was (Wayland1.0 + the useful parts of X11 to have network (WAN) access), then it should be possible to have the toolkits use this..

Wayland - Beyond X (The H)

Posted Feb 16, 2012 13:41 UTC (Thu) by farnz (subscriber, #17727) [Link] (2 responses)

XRender is rather limited in functionality, as compared to OpenGL (only considering the 2D rendering case here). As a result, you end up rendering parts of the toolkit as bitmaps (which you ship over the wire as-is when running networks), when a better API like OpenGL lets you offload work to hardware. Toolkit authors want to produce prettier toolkits than XRender can support (see Clutter for an example of a toolkit that makes use of OpenGL features that just can't be efficiently rendered via XRender).

If they're happy with X11's protocol, they're unlikely to go to the effort of porting to Wayland - why bother coding a completely new backend, when your existing X11 backend will already work with Wayland? Equally, why limit yourself to systems running Wayland if you can work with any systems that run X11?

Note that Wayland is still a success if all it does is act as the construction kit for X11 systems that integrate window management, input handling, and compositing. It doesn't need all the clients to talk native Wayland to have succeeded.

Wayland - Beyond X (The H)

Posted Feb 16, 2012 17:18 UTC (Thu) by renox (guest, #23785) [Link] (1 responses)

> XRender is rather limited in functionality, as compared to OpenGL
Yes but:
1) OpenGL is a moving target
2) You cannot simply send OpenGL commands from the client to the server otherwise any command to allocate something would create a round trip.
So you'd have to have a smart proxy on the client, goto (1).
3) With some luck, OpenGL commands to read pixels from buffers would not be an issue as they're already slow, so they're unlikely to be used.

> If they're happy with X11's protocol, they're unlikely to go to the effort of porting to Wayland

I don't understand your remark: there is work to port GTK and Qt to Wayland, clearly some developers want the faster local rendering that Wayland should provide..

The question is about the remote rendering backend: should the X11 backend be kept forever? Can it be improved for WAN access(*)?
Or is-it possible to replace it with something even better for WAN on top of Wayland?

*:After all, in the long run (if Wayland is ported to other Unix), this should be its main purpose.

Wayland - Beyond X (The H)

Posted Feb 16, 2012 18:42 UTC (Thu) by farnz (subscriber, #17727) [Link]

While OpenGL is a moving target, it is a standard, and what's on offer is extended, not reduced. You can choose a useful subset, available on all interesting platforms, and just use that, while still getting more than XRender offers you. Given that people are moving toolkits to OpenGL, there's clearly a demand for more than XRender offers from the perspective of toolkit authors.

Explaining my remark in more depth; nothing in the Wayland world stops people sticking to X11 indefinitely. If Wayland + XRender (or Wayland + OpenGL 1.4 - the upper limit of indirect GLX) is Good Enough, toolkit authors are unlikely to expend the effort to port to Wayland for no net gain. Given that the only gain from Wayland is to permit you to drop legacy code related to the network-transparent subset of X11, the only sensible reason for toolkits to port over and stop caring about X11 is that they want to do things that they can't do network-transparently in X, either.

Therefore, should toolkits stop bothering to maintain a network-transparent X11 backend, it will be because they're doing things that they couldn't do in network-transparent X11. The benefit of Wayland here is that they have to explicitly decide that "not network transparent" is acceptable - they can't simply be an X11 application that does not work over the network due to DRI2 dependencies.

Thus, the toolkits have to make a decision about remote rendering; is it an irrelevance to their users? Is X11 good enough (in which case they will maintain an X11 backend)? Is there a better way (in which case they will implement it, and possibly kill off Wayland in the process)?

Wayland - Beyond X (The H)

Posted Feb 15, 2012 15:55 UTC (Wed) by nix (subscriber, #2304) [Link]

When you think about it, this is mostly what we have today. Most of what core X does (drawing lines, arcs, rendering fonts and the like) is stuff which would these days go into a toolkit. Obviously the point at which this semantic information is available is the point at which network-transparency should be implemented. If you implement it by throwing huge bitmaps around -- whether in a Wayland compositor, or in present-day X handling current Pango-and-Gtk apps over a network -- you're going to get lousy performance.

Wayland - Beyond X (The H)

Posted Feb 15, 2012 19:10 UTC (Wed) by jedidiah (guest, #20319) [Link] (1 responses)

I suppose you could consider Wayland as a way to rebuild Xorg from the ground up but that's not the way it's being presented. It primarily seems to be a rallying point for X haters. The point really seems to be to abandon X entirely. Lip service given to features missing in Wayland just seem to be an attempt to deflect justifiable criticism.

Wayland - Beyond X (The H)

Posted Feb 17, 2012 9:57 UTC (Fri) by alankila (guest, #47141) [Link]

The ultimate problem here is that getting good local and remote network performance simultaneously is very hard. And we currently sacrifice great deal of local performance through having a middle-management layer such as X, or protocols like GLX, or AIGLX, or whatever.

The idea that clients have direct access to GPU and can render directly with it, and everything else is told to get out of the way will allow superb local performance, but at the cost of removing X-style network transparency because there is no component other than the client and the GPU that even knows what is being rendered. So the ways to get network transparency are:

1) request GPU for textures, try to compress them and send them;
2) use toolkit which actually has the semantic knowledge of what is being displayed and can use some intelligent interface to represent the gui data.

Hence the perfectly reasonable request that toolkits should do the network transparency, because they have the semantic information to actually do a good job.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 10:06 UTC (Tue) by renox (guest, #23785) [Link] (7 responses)

[Do you think the major toolkits will rip out X11 support the day Wayland hits 1.0 in some sort of mass suicide pact?]

Remember that a lot of the toolkit maintenance is made by the desktop projects, which are not especially famous for their backward compatibility..

So it's quite probable that very soon after Wayland is usable (not 1.0 obviously), they'll say use VNC instead of X/FreeNX if you need remote access, we don't support X anymore. It sucks on WAN? Well we don't care.

After all, if developers really cared about remote (WAN) access, FreeNX would have been integrated in the X project one way or the other.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 13:43 UTC (Tue) by nye (subscriber, #51576) [Link] (2 responses)

>After all, if developers really cared about remote (WAN) access, FreeNX would have been integrated in the X project one way or the other.

I'm curious about this: when I tried FreeNX (it was a couple of years ago; maybe that's important) I found it to be *far* worse than x11vnc, which is *just barely* usable over ADSL. X over ssh of course was worlds away from usable - many seconds for minor display updates for example.

Are other people finding FreeNX to be significantly better than x11vnc?

(And of course, they're all atrocious compared to RDP to a Windows machine)

Wayland - Beyond X (The H)

Posted Feb 14, 2012 16:49 UTC (Tue) by drag (guest, #31333) [Link] (1 responses)

> Are other people finding FreeNX to be significantly better than x11vnc?

Yes. Very much so.

Wayland - Beyond X (The H)

Posted Feb 15, 2012 12:11 UTC (Wed) by nye (subscriber, #51576) [Link]

>> Are other people finding FreeNX to be significantly better than x11vnc?

> Yes. Very much so.

Sorry to keep pestering - I'm trying to work out what might be the differences in situations which would explain why people in this thread have such wildly varying experiences:

Do you know in what kind of environment NX shines compared to x11vnc?
I'm mostly thinking of normal, consumer-level ADSL links, so RTT of 50-150ms (where 50ms is on a good day with no load, and 150 is about the worst you're likely to see until bufferbloat kicks in), uplink speed of around 256kbps (768 if you're really lucky), extreme bufferbloat (so RTT can occasionally go up into the seconds if the connection is loaded beyond 90% or so).

In that scenario you have to use so much compression with VNC that it looks awful because the screen's covered with JPEG artefacts, but it's good enough for occasional remote administration at least. RDP to Windows has similarly tolerable performance while looking fine, and both X and FreeNX have been completely unusable for me.

Wayland - Beyond X (The H)

Posted Feb 15, 2012 19:52 UTC (Wed) by jedidiah (guest, #20319) [Link] (3 responses)

Wayland itself is already a "mass suicide pact". If the Wayland developers and proponents are quick to leave us out in the cold then the toolkit developers can do the same. It only requires that the toolkit maintainers have the same disregard for the end users that the Wayland developers do.

Wayland - Beyond X (The H)

Posted Feb 15, 2012 20:01 UTC (Wed) by dlang (guest, #313) [Link] (2 responses)

The arguments that "nobody needs that" from the Wayland people sound very similar to the arguments made by the GNOME people. I could easily see GNOME adopting a Wayland-only approach.

Wayland - Beyond X (The H)

Posted Feb 15, 2012 20:35 UTC (Wed) by rahulsundaram (subscriber, #21946) [Link]

You are again making the mistake of confusing random posters and the position of Wayland developers. "nobody needs that" hasn't been uttered by any Wayland developer.

Wayland - Beyond X (The H)

Posted Feb 15, 2012 23:34 UTC (Wed) by HelloWorld (guest, #56129) [Link]

The arguments that "nobody needs that" from the Wayland people
When did any Wayland developer ever say anything like that?

I use network transparency

Posted Feb 14, 2012 22:07 UTC (Tue) by david.a.wheeler (subscriber, #72896) [Link]

I do run GUI programs remotely, using X (and ssh). So I do depend on network transparency. I'm sure I'm not unique, either.


Copyright © 2026, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds