Wayland and Weston 1.0 released
Posted Oct 24, 2012 4:33 UTC (Wed) by daniels
In reply to: Wayland and Weston 1.0 released
Parent article: Wayland and Weston 1.0 released
The Wayland answer is "nobody wants to do that, and anyone who says they do doesn't know what they are talking about", which doesn't do much to satisfy people's concerns who are using the features that they are so dimissive of.
Oh, for christ's sake. Where has anyone said that? Who?
Wayland already has working network transparency. It's not in a release, but you can see Kristian demoing it in the presentation he gave at XDC. I've seen it with my own eyes; it's pretty cool.
It's a slightly different approach to the way X does it; X forces it in the protocol, and then clients have to go to entertaining lengths to figure out if they can or can't use SHM/GEM/DRI2/etc. Wayland's (to be more accurate, Weston's) remoting support instead implements a proxy compositor and then pipes things over the network making full usage of compression for images, etc. So we've pulled the latency for clients right down to the bare minimum, and we can do world-class video compression without them having to care about it.
So, it's a much much better remoting protocol than X, which is pretty much the worst remote window system protocol anyone's ever seen. No image compression, all the images forced into the same stream as the control messages, which are extremely chatty and round-trip heavy - so it's incredibly sensitive to both latency and throughput. And if you have high/variable latency and low/variable throughput, it's going to be totally, totally unusable.
Please do me a favour in future and do a little background research - even a quick browse through the comments on any other LWN article about Wayland ever would've helped - before you start slagging off free software projects. Thanks.
to post comments)