Burning Man looks like a tentative near-future where robots work the menial crap for us and as we dip our toes in the waters of Culture-like leisure everyone experiments being a hippie and a DIYer artisan and a stoner and a furry and a badass craftsman all at once, and you know what? I'm glad it exists. Bring along the rabbit ears, I'll bring the swiss army knife, and let's see what we can make work.
Oh, good example. Y'all probably know that already, but Star Control 2 is now available freely under an open source license, just under the different name of Ur-Quan Masters due to trademark issues.
I almost want to believe he's deliberately teasing the authorities into increasing the surveillance around the embassy, at a time when that ongoing expense is causing angry murmurs the general public. That would be pretty clever.
Whenever this topic comes up, we end up discussing what publishers really do.
Every time, someone with some knowledge of how the publishing industry works turns up and explains how there is this long road between the author's draft and the book in your hands, made of editing, copy-editing, typesetting, cover design, marketing and more, and the publisher is the truck driver that sees the draft to the end of that road. That's correct.
But I've come to the conclusion that none of that is the one irreplaceable service publishers perform in the system.
All of the above can, to some extent, and for a fee, be performed just as well by independent contractors. (There are great independent editors out there, and aren't we all glad for that.)
The one important thing publishers do is: they take the loss on books that don't earn out.
Now hear me out.
I know how we, Slashdot readers, tend to think about those things. In our minds, if the book doesn't earn out (that is, it brings in less money than the publisher gave the author as an advance), then it's got to be someone's fault, right? Bad writer, bad publisher. Something.
Wrong.
The thing is, a successful book requires a lot of factors. Great writing doesn't suffice. The public is fickle. Yesterday, supernatural romance sold by the truckload, now it doesn't. GRRM was a great writer for decades before you even heard of him. Harry Potter didn't start hitting it big until three or four books into the series. Word of mouth matters, but only after the readership has exceeded a certain critical mass. And until then... someone has to take the loss.
Because, here's the thing. GRRM, Rowling, they're outliers. Many books -- most books, AFAIK -- don't quite earn out. Many deserve to, but don't, because that's not how the world works.
But they still got written, you still read some of them, you still loved some of them, and that only happened because someone, somewhere, was willing to pay an author to keep writing, and take the risk that the great book in their hands may not earn that money back.
And that, friends, is what publishers really do.
Re:30% for publisher? Why have a publisher?
on
Amazon's eBook Math
·
· Score: 1
Goddammit Slashdot. That was "cp book.epub <your device>". That's what I get for not previewing, I guess.
Re:30% for publisher? Why have a publisher?
on
Amazon's eBook Math
·
· Score: 1
The publisher does the funding, editing, copy-editing, typesetting, marketing, legal registration, cover design on the book, and that's not even the one most important role they perform in the system.
I hate to state the obvious, but right now Amazon is attempting to muscle in a 43% markup just to do, basically, cp/publisher/book.epub/customer/. There is no reason I can see to explain this large a margin for digital distribution. So I wouldn't necessarily trust them to keep prices all that low once they get free run of the market.
I've tried a lot of desktops over the years and always returned to KDE as the most able to be useful when I need it to and stay the fuck out of the way the rest of the time. (Unity, despite its reputation, is good at that too.) But the love was no longer really there. Like a favorite old workhorse that you just no longer really ride for the pleasure of it alone.
So I've not kept track of KDE 5 developments, and honestly I expected to be way underwhelmed. It was, after all, supposed to be mainly a port of the same old thing to the new Qt 5.
But I just tried the live CD linked in the article and, uh, whoa. It looks so *tidy*. Full of that orderly neatness that Gnome, for all its faults, has generally been better at than KDE. And I find myself excited for the first time in a long while, and that's a very nice feeling to rediscover.
I really just wish desktops were capable of only turning on the discrete GPU when playing games, and relying on the CPU built-in one the rest of the time. (Or is it possible nowadays and I never found out?)
The open source Gallium3D driver for Southern Island Radeon GPUs has come a LONG way in the recent months. Given a 3.14+ kernel and the soon-to-be-released 10.2 Mesa libs, you can expect performance within 80% of that of the Catalyst driver, and it only keeps getting better. The stability is also pretty good. I love being able to flip smoothly between a full screen game and a chat window or a Web browser.
Turns out we are the Great Ancients from a million years ago that came from the cosmos to seed life. Whatever species ends up evolving there will dig into their past with wonder and trepidation to discover who we were. And then they'll find out about Honey Boo Boo. Ah, to be a fly on the wall...:)
It's a learn-as-you-go sort of deal. Don't start off with the end boss and you'll be okay. I would be pants-pissingly terrified in some of the situations described here, but some of the situations I've been in would make you at least a little queasy, I wager.
As an addendum to the press card thing: if you're working for or with any sort of official organisation that the locals would know of, find their logo, even a crappy small one ripped from a website, print it out on a large piece of paper, US Letter or so, along with the org's name in large bold letters, and tape that to the inside of your windshield. Does not open every door. But does open many doors.
Maybe you've been lucky enough to have that once in a lifetime great teacher. The kind of teacher who somehow explains stuff in such a way that everything makes sense to you; things follow logically from one another and it all seems obvious when he explains it. (And you may not even realize it until he falls sick and the substitute trying to explain the exact same stuff leaves you confused and baffled.)
Elegant code has the same property of apparent obviousness. You read it and just nod because it makes sense and flows logically. There isn't one single way to achieve this, of course. It's not about functional vs. imperative vs. object oriented, but how you employ them for clarity.
Needless to say, such clarity is a very hard property to achieve, and a lifetime of experience will only let you approach it asymptotically. It's still worth the attempt, though.
Thank you for the additional details. You are right -- I meant to make it clear that the Wayland design was thought up by people with some serious experience of the internals and limitations of X, and not a competing team of newcomers, as appears to be assumed all too often. But yes, things aren't as simple as I made them look and there is only a partial overlap between the Wayland devs and the Xorgs devs. Thank you for the correction.
I also agree that Wayland is largely about canning the legacy in order to make current and future needs easier to tackle.
I don't agree with your opinion of the move as a technical choice, though, for three reasons.
1/ Taking X out of the rendering loop does not mean dropping X altogether. It just means that future X servers, when and where they are still needed, will run on top of Wayland. It does deprecate X as the default API, yes. But that's not remotely the same as breaking compatibility.
2/ The comments that Daniel Stone (core Xorg and Wayland dev) made in that oft linked video aren't in agreement with the idea that everything Wayland does can be done on top of X, let alone done well. In his talk, DS mentions e.g. issues with input management when one window wants to grab every input that can't be solved in X.
3/ As a more general philosophical principle, the world moves and everything changes. Everything has a shelf life, up to the universe itself, and there is a point where resisting change for the sake of keeping past things going becomes harmful. And this is the actual reason I've been so active in this thread. Not just because I've got a pretty good hunch that once the dust settles Wayland will largely work better than X. But because I think that we, Slashdotters, Linux users, geeks and nerds, are becoming fearful of change, and that's not a good thing. This, here, is an entire new toy and it opens entire new possibilities! It may break shit and it may be awesome and it will probably be a bit of both. Let's freaking check out the code and play with it! Is this not exactly what we should be about?:)
Have a safe flight, and thank you for the constructive reply!
> Nobody uses Weston so the fact it has RDP support is dick all use to anyone using GNOME or KDE.
Ding ding ding. This, here, is what I think is the main problem with the Wayland ecosystem as it currently exists.
As things currently stand, the Wayland protocol is designed to give compositors a lot of flexibility in what kind of buffers they support, with what capabilities.
The drawback is fragmentation.
So okay, the Unix world at large is not a newcomer to fragmentation issues. But it's still a problem that will have to be addressed.
As I said in another comment, I think that things will probably converge on the common ground of either a de facto standard compositor, or a set of common libraries. Wayland itself will probably ship with a generic, non-optimized implementation for common capabilities like remoting.
But until then, it is an issue, and it would be dishonest not to acknowledge it.
> It will, however suck just as hard as all the other pixel-scrapers.
You are doing the same thing again.
You are asserting, without proof, that Wayland remoting will necessarily have to work as a pixel scrapper.
Wayland is designed to be extensible with regards to supported buffer types. For instance, YUV video buffers were merged into the reference compositor at some point in 2012.
And remoting specific buffer types can be done vastly more efficiently than with a generic pixel scrapper.
For instance, in case of a video buffer, the content can be streamed from the remote app to the local YUV buffer as a lossy video stream without impacting the quality of the other, non-video buffers of the application. (In fact, I think I remember something to that extent being demonstrated somewhere, although I can't find the link, so don't quote me on this.)
Meanwhile, because Xorg's current rendering extensions are not network-transparent, remoting X applications that use those extensions (which will be most of them, these days) does already boil down to filling generic pixel buffers and pushing them down the wire, which works with the approximate efficiency of... a pixel scrapper.
A Wayland stack should therefore be no worse than a current Xorg stack in most cases, and can theoretically be made to work vastly better. (I don't, however, think we are there yet. See below.)
If you're using only pure-X11 apps, though, stick with Xorg for the time being. At least until the Wayland stack supports X buffers sufficiently seamlessly.
> This leads old X11 users to believe they are either liars or incompetent (and hence the lack of trust) because having experienced them all over many years, I know for a fact it is a long way from the worst.
Only if you are willing to make the mental jump of generalizing the one data point of your own limited experience to the entire world and all the use cases that the Xorg developers have to contend with.
If you're gonna give that much weight to single data points, well, I'm afraid that my own long experience of X11, which has most recently involved telecommuting over a VPN and having to use tools like x11vnc and xpra because Xorg remoting on its own works too poorly, neutralizes your own single data point, and we're back to square one.:)
Unless, of course, you are the sort of person who thinks that their own personal experience constitutes the One Overriding Truth, in which case I don't think there is anything whatsoever to be gained for either of us in this discussion.
> The remote windowing is built into the compositor/windowmanager? I really, really hope someone is making a more sensible architecture than that out there
You appear to think yourself an authority on sensible architectures, so I would suggest you prove it with some actual design and implementation. Otherwise, I hope you will forgive me if I don't trust you nearly as much as I do the guys who are doing the actual work based on actual experience with actual issues.:)
That said:
> otherwise we're going to wind up with a lot of deficient compositors or a lot of duplicated code.
YES.
While it could be argued that the many composers/WMs on X already constitutes a lot of duplicated code, I DO agree that the Wayland stack is probably going to be somewhat fragmented for a while, at least until the dust settles.
In practice, I suspect Wayland stacks will end up converging over a shared common ground: probably one display server/composer will emerge as the de facto standard, like Xorg has become the de facto standard of X11 display servers; or Weston itself will evolve into a set of libraries that display servers/composers will be built upon.
But this is only a possible future, and until then, I do think that fragmentation and mismatched composer capabilities is the one big issue with the Wayland budding ecosystem.
Time will tell how big an issue it turns out to be in p
Hi! I would have liked to thank you for taking the time to reply, but I find it difficult to do sincerely, because your comment adds nothing to the discussion, and even detracts from it.
You appear to be trying to imply, with some aggressiveness, that Wayland precludes remoting.
You could have expressed this concern as an interrogation, and I would gladly have tried to share what I understand about the topic.
Instead, no, a bold, unsubstantiated, sarcastic claim with no room for discussion. This is exactly what I meant about the peanut gallery.
Likewise, the fact you refer to the issues the Xorg developers themselves -- you are aware that they are who you refer to as 'the Wayland folks', right? -- have claimed to have with X as 'FUD' makes you look like someone who dismisses contrary opinions instead of addressing them. That really doesn't help the discussion either.
Unless you do, in fact, know better than the Xorg developers themselves. In which case I'm sure you'll step forward to take over the maintenance of Xorg. Won't you?
As to your implied claim, I addressed it in my reply to Uecker below.
I don't personally know the X protocol well enough to comment either way; I can only report on the opinion professed by the X developers themselves.
Here is what I understand are the answers to the points you raise. You'll probably want to check out the talks by Daniel Stone (core Xorg developer) that have been linked elsewhere in the thread, in case I missed something.
Essentially, their opinion is that the X protocol is unsuited to what computers do nowadays. From what I understand, the only task that X11 still performs in current graphic stacks is IPC for the actual rendering extensions, and sadly, IPC is something it's very poor at.
Core X11 is network-transparent by design, but rendering is done through non-core, non-network transparent extensions nowadays. This appears to be a common misunderstanding about the meaning of network-transparent; remote display of application does in fact not require network-transparency because transparency means a lot more than just "remote capable". And Wayland as a protocol is already as remote capable as Xorg because Xorg was already filling buffers remotely and feeding that into SSH connections.
So if you want network-transparency, you'll have to disable all those rendering extensions in your Xorg configuration. But, correct me if I'm wrong, I believe what you really want, is to fire up apps remotely and have them display locally, right? And this has already been implemented in the reference Wayland compositor.
Until then, you will probably not miss the loss of network-transparency because you already lost it in current Xorg servers. I think that this, there, is the number one misunderstanding about both Xorg and Wayland.
You are correct about a window system being more than just about sharing buffers. All the things you mention are being redesigned as part of Wayland with the purpose of fixing issues that the X developers claim were unsolvable with X. (Don't take my word for it, though. Check out those talks.)
So in essence, the X developers think that Wayland stacks will be better than Xorg stacks at everything that Xorg does. Including remoting.
I do actually have one reservation about that general claim, and interestingly, it's one that I haven't seen come up from the aforementioned peanut gallery. But time will tell.
But until then, the X developers think that designing a new API from scratch is more straightforward than monkeypatching the old one into doing the same things. If you sincerely think they are wrong, maybe you'll want to step forward and take over the maintenance of Xorg? I'm sure some people out there would be grateful.
I, for one, am going to trust that they know what they are doing, but you may feel otherwise about that, and that's fine. There's just a "put up or shut up" line there that people who share your opinion seem unwilling to cross, and I think that's worth pointing out.
The Xorg developers got tired of spending their time working around the way X was designed in 1980 (which made sense at the time) to try and make it fit 2010 workloads and hardware.
They started to think about how to do the stuff that actually needs doing in an efficient manner, while removing the roadblocks they currently have to contend with.
Turns out that when you take what Xorg actually does nowadays, streamline the fuck out of it, and take away all the needless obstacles, you end up with a pretty straightforward buffer sharing protocol. They called it Wayland and started to work on an implementation.
And then the countless people in the peanut gallery who obviously know X much better than the X developers beheld the notion and started giving... loud feedback, shall we say. Without ever stepping forward to take over the maintenance of Xorg, mind you.
TL;DR: Xorg developers make what they concluded is the soundest technical choice. People on the Internet lose their shit. Business as usual.
Re:For a dying language Ruby is doing great
on
Ruby 2.1.0 Released
·
· Score: 1
Linux here, works fine for me. Chrome 31. I'd ask "can you try to upgrade to the latest version?", but I wouldn't want to contribute to, you know, Slashdot losing it.:)
Countries like France and UK, yeah, absolutely. Germany... is slightly more touchy about issues pertaining to surveillance and the general topic of totalitarianism, for some reason.
Iceland overthrew its government when said government wouldn't jail bankers. If Iceland says they ain't going to spy on people because fuck that, I would lean toward cautiously trusting them.
Oh, then this is probably my cue to turn up and point you to Mystcraft. Mystcraft is just as the name implies: a mod for Minecraft that adds Myst-like mechanisms for creating and exploring Ages. Yes, it's just about as fantastic as it sounds. (And by God, don't forget to bring a linking book if you don't want to get forever stranded.)
And to answer the question raised by the article, I just spent my lunch break playing Mystcraft. Today, in 2013, 20 years after the release of Myst. So I'd say, pretty relevant indeed.
I think Mir is a case study in how to correctly identify problems and then going about solving them all wrong.
See, the good thing about Wayland is, it does the right thing in having a limited scope. It aims to do one thing and do it well: provide an API for GUI clients to share buffers with a compositor.
And the problem with Wayland is, of course, that... it has a limited scope. Screen management? Input handling? Buffer allocation? "A modern desktop needs all that!" say the Ubuntu devs, and yeah, that's absolutely correct. "That's a client concern," say the Wayland devs, and guess what? From their point of view, that's correct too. (Although Wayland since started working on an input handling API.)
Now, the important thing to realize is, when the Wayland guys say that something is a client concern, as I understand, they don't necessarily mean the GUI applications, no. They mean the compositor.
Meaning that a whole lot of the stuff desktop shells rely on is, in fact, not provided by Wayland itself.
That's where Weston comes in: it's supposed to be an example (a "reference implementation", to use the designated words) of how to write a compositor. But... not necessarily in a way that meets the higher level needs of desktop shells. Unsurprisingly, both KDE and GNOME will be using their own compositors.
So basically, a whole lot of the desktop integration on top of Wayland will be, as it were, left as an exercise to the reader.
With all that in mind, I think the highest outcome end game is somewhat clear: frame-perfect rendering through the Wayland API of Mir-composited KDE/GNOME/Unity clients.
Or in other words, Mir should probably be a set of APIs to handle all the admittedly important desktop integration -- clipboard, multi-screen layout, input and gestures, systray/notification requests... -- with an optional and replaceable compositor thrown in.
All the points of contention that I know of, mainly that Canonical requires server-side buffer allocation (presumably for mobile ARM platforms) where Wayland does it client-side, could have been resolved with some diplomacy and a mutual willingness to reach a satisfactory compromise.
But instead, it looks like the report card is just going to say, "Doesn't play well with others." As usual. What a sad mess and wasted opportunity.
Slashdot Mirror
Burning Man looks like a tentative near-future where robots work the menial crap for us and as we dip our toes in the waters of Culture-like leisure everyone experiments being a hippie and a DIYer artisan and a stoner and a furry and a badass craftsman all at once, and you know what? I'm glad it exists. Bring along the rabbit ears, I'll bring the swiss army knife, and let's see what we can make work.
Oh, good example. Y'all probably know that already, but Star Control 2 is now available freely under an open source license, just under the different name of Ur-Quan Masters due to trademark issues.
I almost want to believe he's deliberately teasing the authorities into increasing the surveillance around the embassy, at a time when that ongoing expense is causing angry murmurs the general public. That would be pretty clever.
Whenever this topic comes up, we end up discussing what publishers really do.
Every time, someone with some knowledge of how the publishing industry works turns up and explains how there is this long road between the author's draft and the book in your hands, made of editing, copy-editing, typesetting, cover design, marketing and more, and the publisher is the truck driver that sees the draft to the end of that road. That's correct.
But I've come to the conclusion that none of that is the one irreplaceable service publishers perform in the system.
All of the above can, to some extent, and for a fee, be performed just as well by independent contractors. (There are great independent editors out there, and aren't we all glad for that.)
The one important thing publishers do is: they take the loss on books that don't earn out.
Now hear me out.
I know how we, Slashdot readers, tend to think about those things. In our minds, if the book doesn't earn out (that is, it brings in less money than the publisher gave the author as an advance), then it's got to be someone's fault, right? Bad writer, bad publisher. Something.
Wrong.
The thing is, a successful book requires a lot of factors. Great writing doesn't suffice. The public is fickle. Yesterday, supernatural romance sold by the truckload, now it doesn't. GRRM was a great writer for decades before you even heard of him. Harry Potter didn't start hitting it big until three or four books into the series. Word of mouth matters, but only after the readership has exceeded a certain critical mass. And until then... someone has to take the loss.
Because, here's the thing. GRRM, Rowling, they're outliers. Many books -- most books, AFAIK -- don't quite earn out. Many deserve to, but don't, because that's not how the world works.
But they still got written, you still read some of them, you still loved some of them, and that only happened because someone, somewhere, was willing to pay an author to keep writing, and take the risk that the great book in their hands may not earn that money back.
And that, friends, is what publishers really do.
Goddammit Slashdot. That was "cp book.epub <your device>". That's what I get for not previewing, I guess.
The publisher does the funding, editing, copy-editing, typesetting, marketing, legal registration, cover design on the book, and that's not even the one most important role they perform in the system.
Amazon does cp book.epub .
Which one here is the middleman exactly?
I hate to state the obvious, but right now Amazon is attempting to muscle in a 43% markup just to do, basically, cp /publisher/book.epub /customer/. There is no reason I can see to explain this large a margin for digital distribution. So I wouldn't necessarily trust them to keep prices all that low once they get free run of the market.
I've tried a lot of desktops over the years and always returned to KDE as the most able to be useful when I need it to and stay the fuck out of the way the rest of the time. (Unity, despite its reputation, is good at that too.) But the love was no longer really there. Like a favorite old workhorse that you just no longer really ride for the pleasure of it alone.
So I've not kept track of KDE 5 developments, and honestly I expected to be way underwhelmed. It was, after all, supposed to be mainly a port of the same old thing to the new Qt 5.
But I just tried the live CD linked in the article and, uh, whoa. It looks so *tidy*. Full of that orderly neatness that Gnome, for all its faults, has generally been better at than KDE. And I find myself excited for the first time in a long while, and that's a very nice feeling to rediscover.
I really just wish desktops were capable of only turning on the discrete GPU when playing games, and relying on the CPU built-in one the rest of the time. (Or is it possible nowadays and I never found out?)
The open source Gallium3D driver for Southern Island Radeon GPUs has come a LONG way in the recent months. Given a 3.14+ kernel and the soon-to-be-released 10.2 Mesa libs, you can expect performance within 80% of that of the Catalyst driver, and it only keeps getting better. The stability is also pretty good. I love being able to flip smoothly between a full screen game and a chat window or a Web browser.
Turns out we are the Great Ancients from a million years ago that came from the cosmos to seed life. Whatever species ends up evolving there will dig into their past with wonder and trepidation to discover who we were. And then they'll find out about Honey Boo Boo. Ah, to be a fly on the wall... :)
It's a learn-as-you-go sort of deal. Don't start off with the end boss and you'll be okay. I would be pants-pissingly terrified in some of the situations described here, but some of the situations I've been in would make you at least a little queasy, I wager.
As an addendum to the press card thing: if you're working for or with any sort of official organisation that the locals would know of, find their logo, even a crappy small one ripped from a website, print it out on a large piece of paper, US Letter or so, along with the org's name in large bold letters, and tape that to the inside of your windshield. Does not open every door. But does open many doors.
Maybe you've been lucky enough to have that once in a lifetime great teacher. The kind of teacher who somehow explains stuff in such a way that everything makes sense to you; things follow logically from one another and it all seems obvious when he explains it. (And you may not even realize it until he falls sick and the substitute trying to explain the exact same stuff leaves you confused and baffled.)
Elegant code has the same property of apparent obviousness. You read it and just nod because it makes sense and flows logically. There isn't one single way to achieve this, of course. It's not about functional vs. imperative vs. object oriented, but how you employ them for clarity.
Needless to say, such clarity is a very hard property to achieve, and a lifetime of experience will only let you approach it asymptotically. It's still worth the attempt, though.
Hi,
Thank you for the additional details. You are right -- I meant to make it clear that the Wayland design was thought up by people with some serious experience of the internals and limitations of X, and not a competing team of newcomers, as appears to be assumed all too often. But yes, things aren't as simple as I made them look and there is only a partial overlap between the Wayland devs and the Xorgs devs. Thank you for the correction.
I also agree that Wayland is largely about canning the legacy in order to make current and future needs easier to tackle.
I don't agree with your opinion of the move as a technical choice, though, for three reasons.
1/ Taking X out of the rendering loop does not mean dropping X altogether. It just means that future X servers, when and where they are still needed, will run on top of Wayland. It does deprecate X as the default API, yes. But that's not remotely the same as breaking compatibility.
2/ The comments that Daniel Stone (core Xorg and Wayland dev) made in that oft linked video aren't in agreement with the idea that everything Wayland does can be done on top of X, let alone done well. In his talk, DS mentions e.g. issues with input management when one window wants to grab every input that can't be solved in X.
3/ As a more general philosophical principle, the world moves and everything changes. Everything has a shelf life, up to the universe itself, and there is a point where resisting change for the sake of keeping past things going becomes harmful. And this is the actual reason I've been so active in this thread. Not just because I've got a pretty good hunch that once the dust settles Wayland will largely work better than X. But because I think that we, Slashdotters, Linux users, geeks and nerds, are becoming fearful of change, and that's not a good thing. This, here, is an entire new toy and it opens entire new possibilities! It may break shit and it may be awesome and it will probably be a bit of both. Let's freaking check out the code and play with it! Is this not exactly what we should be about? :)
Have a safe flight, and thank you for the constructive reply!
> Nobody uses Weston so the fact it has RDP support is dick all use to anyone using GNOME or KDE.
Ding ding ding. This, here, is what I think is the main problem with the Wayland ecosystem as it currently exists.
As things currently stand, the Wayland protocol is designed to give compositors a lot of flexibility in what kind of buffers they support, with what capabilities.
The drawback is fragmentation.
So okay, the Unix world at large is not a newcomer to fragmentation issues. But it's still a problem that will have to be addressed.
As I said in another comment, I think that things will probably converge on the common ground of either a de facto standard compositor, or a set of common libraries. Wayland itself will probably ship with a generic, non-optimized implementation for common capabilities like remoting.
But until then, it is an issue, and it would be dishonest not to acknowledge it.
> It will, however suck just as hard as all the other pixel-scrapers.
You are doing the same thing again.
You are asserting, without proof, that Wayland remoting will necessarily have to work as a pixel scrapper.
Wayland is designed to be extensible with regards to supported buffer types. For instance, YUV video buffers were merged into the reference compositor at some point in 2012.
And remoting specific buffer types can be done vastly more efficiently than with a generic pixel scrapper.
For instance, in case of a video buffer, the content can be streamed from the remote app to the local YUV buffer as a lossy video stream without impacting the quality of the other, non-video buffers of the application. (In fact, I think I remember something to that extent being demonstrated somewhere, although I can't find the link, so don't quote me on this.)
Meanwhile, because Xorg's current rendering extensions are not network-transparent, remoting X applications that use those extensions (which will be most of them, these days) does already boil down to filling generic pixel buffers and pushing them down the wire, which works with the approximate efficiency of... a pixel scrapper.
A Wayland stack should therefore be no worse than a current Xorg stack in most cases, and can theoretically be made to work vastly better. (I don't, however, think we are there yet. See below.)
If you're using only pure-X11 apps, though, stick with Xorg for the time being. At least until the Wayland stack supports X buffers sufficiently seamlessly.
> This leads old X11 users to believe they are either liars or incompetent (and hence the lack of trust) because having experienced them all over many years, I know for a fact it is a long way from the worst.
Only if you are willing to make the mental jump of generalizing the one data point of your own limited experience to the entire world and all the use cases that the Xorg developers have to contend with.
If you're gonna give that much weight to single data points, well, I'm afraid that my own long experience of X11, which has most recently involved telecommuting over a VPN and having to use tools like x11vnc and xpra because Xorg remoting on its own works too poorly, neutralizes your own single data point, and we're back to square one. :)
Unless, of course, you are the sort of person who thinks that their own personal experience constitutes the One Overriding Truth, in which case I don't think there is anything whatsoever to be gained for either of us in this discussion.
> The remote windowing is built into the compositor/windowmanager? I really, really hope someone is making a more sensible architecture than that out there
You appear to think yourself an authority on sensible architectures, so I would suggest you prove it with some actual design and implementation. Otherwise, I hope you will forgive me if I don't trust you nearly as much as I do the guys who are doing the actual work based on actual experience with actual issues. :)
That said:
> otherwise we're going to wind up with a lot of deficient compositors or a lot of duplicated code.
YES.
While it could be argued that the many composers/WMs on X already constitutes a lot of duplicated code, I DO agree that the Wayland stack is probably going to be somewhat fragmented for a while, at least until the dust settles.
In practice, I suspect Wayland stacks will end up converging over a shared common ground: probably one display server/composer will emerge as the de facto standard, like Xorg has become the de facto standard of X11 display servers; or Weston itself will evolve into a set of libraries that display servers/composers will be built upon.
But this is only a possible future, and until then, I do think that fragmentation and mismatched composer capabilities is the one big issue with the Wayland budding ecosystem.
Time will tell how big an issue it turns out to be in p
Hi! I would have liked to thank you for taking the time to reply, but I find it difficult to do sincerely, because your comment adds nothing to the discussion, and even detracts from it.
You appear to be trying to imply, with some aggressiveness, that Wayland precludes remoting.
You could have expressed this concern as an interrogation, and I would gladly have tried to share what I understand about the topic.
Instead, no, a bold, unsubstantiated, sarcastic claim with no room for discussion. This is exactly what I meant about the peanut gallery.
Likewise, the fact you refer to the issues the Xorg developers themselves -- you are aware that they are who you refer to as 'the Wayland folks', right? -- have claimed to have with X as 'FUD' makes you look like someone who dismisses contrary opinions instead of addressing them. That really doesn't help the discussion either.
Unless you do, in fact, know better than the Xorg developers themselves. In which case I'm sure you'll step forward to take over the maintenance of Xorg. Won't you?
As to your implied claim, I addressed it in my reply to Uecker below.
Hi! Thank you for taking the time to reply.
I don't personally know the X protocol well enough to comment either way; I can only report on the opinion professed by the X developers themselves.
Here is what I understand are the answers to the points you raise. You'll probably want to check out the talks by Daniel Stone (core Xorg developer) that have been linked elsewhere in the thread, in case I missed something.
Essentially, their opinion is that the X protocol is unsuited to what computers do nowadays. From what I understand, the only task that X11 still performs in current graphic stacks is IPC for the actual rendering extensions, and sadly, IPC is something it's very poor at.
Core X11 is network-transparent by design, but rendering is done through non-core, non-network transparent extensions nowadays. This appears to be a common misunderstanding about the meaning of network-transparent; remote display of application does in fact not require network-transparency because transparency means a lot more than just "remote capable". And Wayland as a protocol is already as remote capable as Xorg because Xorg was already filling buffers remotely and feeding that into SSH connections.
So if you want network-transparency, you'll have to disable all those rendering extensions in your Xorg configuration. But, correct me if I'm wrong, I believe what you really want, is to fire up apps remotely and have them display locally, right? And this has already been implemented in the reference Wayland compositor.
Until then, you will probably not miss the loss of network-transparency because you already lost it in current Xorg servers. I think that this, there, is the number one misunderstanding about both Xorg and Wayland.
You are correct about a window system being more than just about sharing buffers. All the things you mention are being redesigned as part of Wayland with the purpose of fixing issues that the X developers claim were unsolvable with X. (Don't take my word for it, though. Check out those talks.)
So in essence, the X developers think that Wayland stacks will be better than Xorg stacks at everything that Xorg does. Including remoting.
I do actually have one reservation about that general claim, and interestingly, it's one that I haven't seen come up from the aforementioned peanut gallery. But time will tell.
But until then, the X developers think that designing a new API from scratch is more straightforward than monkeypatching the old one into doing the same things. If you sincerely think they are wrong, maybe you'll want to step forward and take over the maintenance of Xorg? I'm sure some people out there would be grateful.
I, for one, am going to trust that they know what they are doing, but you may feel otherwise about that, and that's fine. There's just a "put up or shut up" line there that people who share your opinion seem unwilling to cross, and I think that's worth pointing out.
The story so far in a nutshell:
The Xorg developers got tired of spending their time working around the way X was designed in 1980 (which made sense at the time) to try and make it fit 2010 workloads and hardware.
They started to think about how to do the stuff that actually needs doing in an efficient manner, while removing the roadblocks they currently have to contend with.
Turns out that when you take what Xorg actually does nowadays, streamline the fuck out of it, and take away all the needless obstacles, you end up with a pretty straightforward buffer sharing protocol. They called it Wayland and started to work on an implementation.
And then the countless people in the peanut gallery who obviously know X much better than the X developers beheld the notion and started giving... loud feedback, shall we say. Without ever stepping forward to take over the maintenance of Xorg, mind you.
TL;DR: Xorg developers make what they concluded is the soundest technical choice. People on the Internet lose their shit. Business as usual.
It was but a flesh wound!
Linux here, works fine for me. Chrome 31. I'd ask "can you try to upgrade to the latest version?", but I wouldn't want to contribute to, you know, Slashdot losing it. :)
The real disturbing moment was when I rewatched the movie a few years after 9/11 and realized just how much it had anticipated correctly.
Does it still count as satire when it's so spot on?
Countries like France and UK, yeah, absolutely. Germany... is slightly more touchy about issues pertaining to surveillance and the general topic of totalitarianism, for some reason.
Iceland overthrew its government when said government wouldn't jail bankers. If Iceland says they ain't going to spy on people because fuck that, I would lean toward cautiously trusting them.
Oh, then this is probably my cue to turn up and point you to Mystcraft. Mystcraft is just as the name implies: a mod for Minecraft that adds Myst-like mechanisms for creating and exploring Ages. Yes, it's just about as fantastic as it sounds. (And by God, don't forget to bring a linking book if you don't want to get forever stranded.)
And to answer the question raised by the article, I just spent my lunch break playing Mystcraft. Today, in 2013, 20 years after the release of Myst. So I'd say, pretty relevant indeed.
I think Mir is a case study in how to correctly identify problems and then going about solving them all wrong.
See, the good thing about Wayland is, it does the right thing in having a limited scope. It aims to do one thing and do it well: provide an API for GUI clients to share buffers with a compositor.
And the problem with Wayland is, of course, that... it has a limited scope. Screen management? Input handling? Buffer allocation? "A modern desktop needs all that!" say the Ubuntu devs, and yeah, that's absolutely correct. "That's a client concern," say the Wayland devs, and guess what? From their point of view, that's correct too. (Although Wayland since started working on an input handling API.)
Now, the important thing to realize is, when the Wayland guys say that something is a client concern, as I understand, they don't necessarily mean the GUI applications, no. They mean the compositor.
Meaning that a whole lot of the stuff desktop shells rely on is, in fact, not provided by Wayland itself.
That's where Weston comes in: it's supposed to be an example (a "reference implementation", to use the designated words) of how to write a compositor. But... not necessarily in a way that meets the higher level needs of desktop shells. Unsurprisingly, both KDE and GNOME will be using their own compositors.
So basically, a whole lot of the desktop integration on top of Wayland will be, as it were, left as an exercise to the reader.
With all that in mind, I think the highest outcome end game is somewhat clear: frame-perfect rendering through the Wayland API of Mir-composited KDE/GNOME/Unity clients.
Or in other words, Mir should probably be a set of APIs to handle all the admittedly important desktop integration -- clipboard, multi-screen layout, input and gestures, systray/notification requests... -- with an optional and replaceable compositor thrown in.
All the points of contention that I know of, mainly that Canonical requires server-side buffer allocation (presumably for mobile ARM platforms) where Wayland does it client-side, could have been resolved with some diplomacy and a mutual willingness to reach a satisfactory compromise.
But instead, it looks like the report card is just going to say, "Doesn't play well with others." As usual. What a sad mess and wasted opportunity.