LWN: Comments on "The 2013 "State of X.Org" report"

The 2013 "State of X.Org" report

oak — Wed, 27 Feb 2013 20:12:37 +0000

Making messages short and concise can require a lot more time and attention than quick stream-of-consciousness typing.

If writer doesn't feel obliged to do that, I don't feel much obliged to waste my time trying to understand him.

“I didn't have time to write a short letter, so I wrote a long one instead.”
-- Mark Twain

The 2013 "State of X.Org" report

nix — Wed, 27 Feb 2013 18:03:53 +0000

I don't understand how you could expect the X server to 'imitate' Linux's architecture of a complicated core and relatively simple drivers. An X server simply does a different job: it's not going to have that architecture any more than, say, a compiler would. You can't "just" totally rearchitect software without reimplementing it completely anyway (which is, surprise, what Wayland is doing).

The 2013 "State of X.Org" report

mmarq — Tue, 26 Feb 2013 22:34:55 +0000

i see... them it should be everybody happy, and no need to run clumsy hacks of server on top of server.

Basic it may be but will be more than enough as you say, since GKT/Qt etc already runs natively on wayland then.

Who ever wants the new fashion, instals wayland, whoever wants to stay using NX installs X... no promiscuity/bickering please, no need at all, just chance the WM. And then the DM login could have 4 desktop options... or more options... Gonme wayland, Gnome X, KDE wayland KDE x... and all the applications will be transparent to this

Good "suggestions"... no ?

The 2013 "State of X.Org" report

beagnach — Tue, 26 Feb 2013 22:24:44 +0000

I would suggest posting fewer comments with less words - it's very hard to understand what you are trying to say because you are very verbose.

The 2013 "State of X.Org" report

mmarq — Tue, 26 Feb 2013 22:06:22 +0000

> assuming you can get anyone to find the time to pay attention to you!

there is a folks saying in my country full of wisdom (IMO);

" with the proper housing, resilience and persistence, there isn't lacking than doesn't end up in festival "

The 2013 "State of X.Org" report

neilbrown — Tue, 26 Feb 2013 22:02:02 +0000

> will imply changing every single one X app out there.

Maybe it depends on what you mean by an "X app".
You shouldn't need to change most Gnome and other GTK apps. GTK already has basic Wayland support.
You shouldn't need to change most KDE and other Qt apps. Qt already has basic Wayland support.
So all these should run natively on wayland.
I suspect TK and wxWidgets will also provide native support if Wayland shows any sign of success. Maybe they already do.

If those are "X apps", then you don't need to change every "X app".
If they are not "X apps", then ... how many "X apps" do you use? xclock? xload? Certainly there are some, but they shouldn't be a big barrier to change, assuming the change brings any value to makes the cost worth it (which I think it will).

The 2013 "State of X.Org" report

mmarq — Tue, 26 Feb 2013 21:51:51 +0000

the kinder reception is great then... and perhaps a point in front of Linux kernel then.

But why not "imitate" the rest ? .. perhaps it could have more than 1 point in front of linux kernel, imitating its "modus operandi".

The 2013 "State of X.Org" report

nix — Tue, 26 Feb 2013 21:37:21 +0000

is the X system more complex than Linux kernel ?

It's more complex than the drivers which most people work on, I'd say, and there isn't really an analogue of the driver layer in the X server: it's a complex and old and involuted and rather brittle hairball in many ways, and like the kernel it too has ways in which you can make mistakes which don't show up until it is too late to fix them (mostly mistakes in extension protocol design).

perhaps X is not easy to get into because its too "closed"... perhaps it doesn't have the proper versioning infrastructure for patch verification discussion and applying

Whatever you mean by 'closed', it is not clarified by your 'proper versioning infrastructure' comment, which simply makes no sense. Lack of what could be called 'review bandwidth' is an issue, but that's a human problem, not a problem that can be solved by 'proper infrastructure' unless you consider skilled X hackers to be 'infrastructure'.

X is not notably unapproachable as free software projects go. You'll probably get a much kinder reception on the X list than on the linux-kernel list -- assuming you can get anyone to find the time to pay attention to you!

The 2013 "State of X.Org" report

mmarq — Tue, 26 Feb 2013 21:18:06 +0000

> I did one suggestion: don't tell people (especially F/OSS people) what they should be doing. Just that. Instead of trying to give orders to people, I suggested you *do* whatever you are trying to order others to do.

Oh! now i understand why.

Sorry if i offended someone with a more "aggressive" tone. Giving orders never was the intend. The more aggressive "outlook" is because i still can't understand why this replacement "endorsement" that will imply changing every single one X app out there... as if must be this and then will be that for sure and no other choice...

The idea is not being a "wet blanket" for wayland, but reality can be very cruel... its not me "the problem" gentlemen, its the billions (yes with a B) of *potential end-users* out there, its not me who you have to listen, convince and put up with, its them. Yes the Ultra-Mobile side can be a refuge, its like an appliance, as if it is this, and take it or live it 'o' end-user... but lets be realistic, none of this projects will ever replace Android, Mac, and or Windows that is growing pushed by multi-billions at the front, anytime soon (if ever).

That is, if you gentleman want to have chances of "mass adoption" with a new project for the client, then the "desktop" side must be the first target, and this may already include Mobile from tablets to laptops.(IMHO not an order)

Those can be a very though audience just for starters... i have some experience, let me tell you... they want all their apps up and running, with no bugs or crashes, they want great looks, and they want all their hardware supported, and this including upto 4 GPU card in a system if it must.

And short the list may seem, those are worst than orders. You just can't send them to slash*t or something, telling them to go away or do the code themselves... 99.9% percent of them NEVER did a line of coding in their lives, those suggestions would seem worst than an insult, they just will not only don't understand, they will stay with Windows/Mac and smear and insult to kingdom come, not only a display system (what ever that may be) but *ALL* of Linux, the good the bad and all in between.

Perhaps for "mass adoption intents" on the desktop/client sector, the all Linux infrastructure will need a PR front... just to try to isolate the devs...

X has almost all of the apps there already, X can have great looks... the all thing lacks yet the proper hardware support that those ignorant massive masses would be complaining about... wayland will get to start allover, and even if it doesn't, those end-users couldn't care less about code design style complexity or language (as long it works well and has no bugs).

The 2013 "State of X.Org" report

mmarq — Tue, 26 Feb 2013 20:13:12 +0000

Ooops!

that is a "suggestion" not an order ... ok ?

The 2013 "State of X.Org" report

mmarq — Tue, 26 Feb 2013 20:01:42 +0000

> The OP was wanting a way to get his feet wet in the world of Linux graphics / X (that's not an easy world to get into, due to the massive complexity of them).

lol... i understood the "suggestion"... but is the X system more complex than Linux kernel ?

Not by shadows. Perhaps i'm not getting why... but it seems to me that perhaps X is not easy to get into because its too "closed"... perhaps it doesn't have the proper versioning infrastructure for patch verification discussion and applying...

That is what i was trying to say, and it seems to me quite readable and understandable. Follow the Linux kernel example, that AFAIK never stopped from being improved and most important never stopped from being ABI compatible since the series 2.

The 2013 "State of X.Org" report

hummassa — Tue, 26 Feb 2013 17:46:06 +0000

> All that "personal attack" because

Nope, no personal attacks at all.

Let's review my words:

>> Ok, I feel like I'm just feeding a troll here, but here's one suggestion: You can do that. Feel free to do just that, and let others work on what they want (hint: they will anyway).

I did one suggestion: don't tell people (especially F/OSS people) what they should be doing. Just that. Instead of trying to give orders to people, I suggested you *do* whatever you are trying to order others to do.

>> Another suggestion: proofread what you say (what's up with all the ellipses?) and after doing that, read it aloud before posting. If you do that, it is possible that you won't come across as insane (you _are_ coming across as a person with deep psychiatric trouble, do you know that?). Feel free to ignore this suggestion and eventually you'll end up in everyone's killfile.

No personal attack here, too. I am criticizing, and quite constructively (for me at least, and taking into consideration the sheer number of posts you did lately versus their quality), the quality and understandability of the text you are writing. I explained that you are putting a lot of ellipsis, and that the flow of your text is awkward. And that people (me included) that read your text either cannot understand your arguments or simply think you are crazy. Notice I did *not* affirm you are crazy. I simply don't know if you are -- like you don't know if I am crazy or not. Just that your text makes people think that.

Finally,

> lol... i think further comments are not necessary... perhaps if you shed all that "hatred"...

I don't hate you. Or anyone else here on LWN, at least not yet! Nobody gave me cause for that. I am passionate about free software, and while I do subscribe to some of the more "radical" Free Software ideologies, I don't think people that make or use proprietary software are evil.

It is my opinion that your texts are just incomprehensible for me, and I am suggesting that you change that, for my benefit.

It is also my opinion that your texts are so incoherent that they make people have a bad image of you, and I am suggesting you change that, for YOUR benefit.

Just that. Peace, and HTH.

The 2013 "State of X.Org" report

Kit — Tue, 26 Feb 2013 17:09:05 +0000

No, he is kinda' right... your posts are very, very hard to read. It has nothing to do with agreeing/disagreeing, and everything to do with the flow of the words.

As for the
> > Why not "improve substantially" what is already there ?

You didn't follow what dgm wrote, what he was responding to, and why he suggested that. The OP was wanting a way to get his feet wet in the world of Linux graphics / X (that's not an easy world to get into, due to the massive complexity of them).

dgm's response was suggesting a simple 'toy' project for the OP to do. It wasn't intended to be directly fruitful (hence 'toy'), but to be a learning experience for the OP. That way, in the future, he'll have an easier time working on "what is already there". GPUs are absurdly complicated pieces of hardware ('absurdly' is very likely a massive understatement), most people can't just dive into working with them and expect to be at all productive anytime soon.

The 2013 "State of X.Org" report

mpr22 — Tue, 26 Feb 2013 16:12:04 +0000

GPL / MIT license compatibility is one-way. A GPL project can incorporate MIT-licensed code, but MIT-licensed projects can't incorporate GPL code - and the X.org developers deliberately chose the MIT license.

The 2013 "State of X.Org" report

drago01 — Tue, 26 Feb 2013 16:11:27 +0000

Not in a sense that you can copy GPL code into it and still retain its licence. That's not how the GPL work. (IANAL).

The 2013 "State of X.Org" report

mmarq — Tue, 26 Feb 2013 15:48:39 +0000

And isn't compatible ?

The 2013 "State of X.Org" report

mmarq — Tue, 26 Feb 2013 15:47:16 +0000

All that "personal attack" because

> Why not "improve substantially" what is already there ?

lol... i think further comments are not necessary... perhaps if you shed all that "hatred"...

The 2013 "State of X.Org" report

drago01 — Tue, 26 Feb 2013 11:32:27 +0000

> Why not merge NX(basically all of it is GPL) with X ?

X is not GPL.

The 2013 "State of X.Org" report

hummassa — Tue, 26 Feb 2013 11:10:38 +0000

> Why not "improve substantially" what is already there ?

Ok, I feel like I'm just feeding a troll here, but here's one suggestion: You can do that. Feel free to do just that, and let others work on what they want (hint: they will anyway).

Another suggestion: proofread what you say (what's up with all the ellipses?) and after doing that, read it aloud before posting. If you do that, it is possible that you won't come across as insane (you _are_ coming across as a person with deep psychiatric trouble, do you know that?). Feel free to ignore this suggestion and eventually you'll end up in everyone's killfile.

The 2013 "State of X.Org" report

mmarq — Tue, 26 Feb 2013 08:42:28 +0000

"" Building on previous prognostications, I quite confidently
predict that 2013 will be the Year of Mobile Wayland. I'm
looking forward to it. ""

Doubt!.. well it depends, if it remains quite simple and trivial and that fulfills the requirements, it might have good chances for "Mobile", but that is all of it.

In the future with even entry level superphones with GPUs 5 to 6x more powerful than today... and video engines and "compute" capabiliites (c'mon, its been announced *already* mobile GPUs by everybody -nVidia, Imagination, ARM, AMD- 5 to 6x more powerful than today... and most of laptops could have 2 of them... ), chances are that X reveals itself more useful than Wayland even for mobile...

i think Android will remain *the* reference in this field... an Android will be using Wayland or not ?

The 2013 "State of X.Org" report

mmarq — Tue, 26 Feb 2013 08:22:01 +0000

Oh! very honestly i can agree with most of what you write "dgm"... but isn't one already there ?

Why not "improve substantially" what is already there ? ... "basic" and "incomplete", is already this particular field of "display/graphics" full of!..

Isn't a popular saying full of wisdom; "union makes the strength" ..right ?

Why not merge NX(basically all of it is GPL) with X ? ... why not a "multithreading", "multi GPU aware" X system ?

(can't find anything more enthusiasm/interest giving, full of life and challenging, than this perspective ... only it might need a lot of "union & cooperation", its not for a single mind, for a single individual glory... c'mon it could be perhaps 1/4 1/5 the size of Linux kernel, and this last project is prove that such things are possible)

The 2013 "State of X.Org" report

dgm — Mon, 25 Feb 2013 09:16:05 +0000

Have you offered yourself for helping on small tasks, like proof-reading drafts? Also, one simple way to get into the code is to write a sample toy driver for the hardware you have, a basic VGA driver for instance.

The 2013 "State of X.Org" report

skitching — Sun, 24 Feb 2013 16:21:37 +0000

A shame about the Book Sprints; I was hoping for some better results there. Quite how anyone becomes an X or linux graphics driver developer is a mystery to me, given the vast complexity of the code and the serious lack of high-level documentation...