In the interests of balance (or, Good job, Canonical!)

Source

Yes, don't fall out of your chairs yet. Here's Canonical being rather decent about upstreaming. Multitouch seems like exactly the kind of thing Canonical ought to be working on, and they're taking this upstream at the right time and with the right approach, AFAICS. That's exactly what I was talking about in this other post. I can't tell whether the proposal is awesome or crack, I'll leave that to the experts. But the process here looks awesome. So nice job with that one.

Comments

Jim Campbell wrote on 2010-08-17 20:58:
Better, yes. It does look like Canonical will require copyright assignment for any patches to their work, though. I trust Canonical enough to handle this fairly, but we can see what happens when companies get bought and the new corporate overlords decide to handle things differently (ala Sun -> Oracle, of course).
adamw wrote on 2010-08-17 21:15:
If it becomes part of upstream X, as is proposed in that mail, it'd be maintained within X.org's RCS and I don't believe would require copyright assignment (I don't think X.org does this).
Jef Spaleta wrote on 2010-08-18 01:24:
Adam... only part of it becomes part of upstream X. The libraries that sit above the X extension and that application developers will be encouraged to use are copyright assigned to Canonical making them very unlikely to be accepted by GNOME upstream as deps. So the evince patches that use these libraries that Canonical is spinning up to showcase this work will most likely not be acceptable as upstream contributions because of the dependency on the libraries that will continue to require copyright assignment. The end game here is that like libappindicator ...the functionality that grail and gies libraries are going to end up being re-implemented as libraries that do not require copyright assignment in order to get the functionality upstream in GNOME. It's not all roses. The copyright assignment requirements on the convenience library layer wedged between X and applications is going to retard adoption of this specific implementation. The copyright assignment policy is really unfortunate and is going to keep reasonable good to excellent engineering work from being incorporated in upstream frameworks. -jef
adamw wrote on 2010-08-18 03:19:
jef: sure. honestly, though, I agree that copyright assignment is not the best policy, but in the grand scheme of things it's fairly small beans. and regardless of whatever Canonical builds on top of the extension, the fact that they're properly contributing the extension (which is really useful work that we need) to X.org shouldn't be discounted.
Juanjo wrote on 2010-08-18 07:02:
I don't think xorg license it's that permissive by accident. If they want to push that spec for a X extension, copyright assignment it's definitely a bad idea for the reference implementation. Anyway, I can't see the benefit for copyright assignment here, do you? Dual licensing?
Jef Spaleta wrote on 2010-08-18 18:25:
Jaunjo, To be clear.. the X extention spec is not copyright assigned to canonical. Like adam has said..they are doing the X extension upstream communication in the right way. They have a toy implementation out in the wild for people to play with..they have acknowledged that a final specification can and will be incompatible with the toy implementation they are shipping in the next Ubuntu release and are willing rework the higher levels of their touch software stack to conform to the final implementation as adopted by upstream X.org. The copyright assignment kicks in a level higher up the softare stack. The libraries geis and grial which sit on top of the X extention that Ubuntu is shipping has an assignment requirement. I would encourage you to read: http://lists.x.org/archives/xorg-devel/2010-August/012048.html Canonical expects the application developers to use API exposed in those libraries not the X extension directly. That is a very wise move and anticipates the need for the X extension to change on upstream review. The problem with the copyright assignment of those libraries is that its going to retard adoption of the gies and grail APIs. Once an official X extension is in place in upstream X, other library API implementations will spring up to provide this functionality in the scope of existing frameworks like Qt and GTK+ that applications can hopefully get for free as part of using the framework without copyright assignment issues. Two steps forward, one step back... but it is progress none-the-less. -jef
Juanjo wrote on 2010-08-19 06:32:
Good pointer, thanks Jef.
Demian wrote on 2010-08-25 10:09:
Nice, multitouch on the way. That, along with windicators and a purple theme, is all Ubuntu needs to fix bug#1.