Wayland - Beyond X (The H)

Posted Feb 14, 2012 9:15 UTC (Tue) by janpla (guest, #11093)
Parent article: Wayland - Beyond X (The H)

Hmm, as far as I can see the argument goes something like "Wayland is necessary because X is rubbish, because X can do too much".

Somehow that argument fails to convince - to my mind, the problem of "too many options" can be solved fairly easily with a sensible set of defaults and/or a good configuration tool, whereas it is a lot harder to solve the problem of features missing because the tool just doesn't support it.

to post comments

Wayland - Beyond X (The H)

Posted Feb 14, 2012 9:56 UTC (Tue) by epa (subscriber, #39769) [Link] (3 responses)

The problem of "too many options" bites you much harder if you are an implementer. You have to add and test support for all the options and all their weird combinations, even those that are hardly used any more or those that could easily be replaced with some simpler feature. Keith P. has first-hand experience of this.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 13:25 UTC (Tue) by sorpigal (subscriber, #36106) [Link] (2 responses)

It would be one thing if wayland were a "better X" - meaning exactly like X in all the ways we like but with a deliberate break of backwards compatibility to get rid of cruft. But it's not.

It seems like ultimately Wayland will be nothing more than a new subsystem used by X and that nothing else will change. I can't forsee getting rid of X any time soon for a replacement that isn't a lot like X, but I can see swapping out the guts of X and modernizing them. That's okay.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 14:05 UTC (Tue) by AndreE (guest, #60148) [Link]

"Modernizing" X requires gutting and reworking much of it's original architecture. We don't need a "better" X, we need "better than X". The actual article makes this pretty clear.

Wayland - Beyond X (The H)

Posted Feb 15, 2012 9:23 UTC (Wed) by jku (subscriber, #42379) [Link]

That is the SVN strategy: "be a better CVS". At the time it sounded like a fairly good idea. Nowadays, I'm sure I'm not the only one who only touches SVN when it's absolutely necessary, and would never choose it for their own projects.

Better X is not enough.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 12:30 UTC (Tue) by nim-nim (subscriber, #34454) [Link]

There are a lot of legacy limitations in X, not all of which can be hacked around as people tried this last decade (for example, any recent keyboard will have keys the kernel recognize but X doesn't, because of protocol limitations)

I guess it will go the same way as the xfd font backend ten years ago: x devs get so fed up with legacy problems they move to something else, and no one steps up to do anything but maintain the old cruft in life-support no-enhancements-or-fixes mode. And apps that do not follow slowlink sink into obscurity.

Wayland - Beyond X (The H)

Posted Feb 14, 2012 17:42 UTC (Tue) by daglwn (guest, #65432) [Link]

Oh, I think the case for Wayland is very good. The X code base is a mess and a good large percent of it really isn't used. It's a maintenance nightmare.

However, nothing I've read indicates that the Wayland people understand just how many people depend on remote X or some equivalent. If they provide an equivalent that performs as well or better and is just as easy to use (that means in particular no setup required on the remote machine), I'm all for it. If not, Wayland is next to useless for me.