|
|
Subscribe / Log in / New account

Multitouch support for Ubuntu 10.10

Canonical has announced the release of uTouch 1.0, a multitouch/gesture stack which will be shipped with the upcoming 10.10 release. "With Ubuntu 10.10 (the Maverick Meerkat), users and developers will have an end-to-end touch-screen framework — from the kernel all the way through to applications. Our multi-touch team has worked closely with the Linux kernel and X.org communities to improve drivers, add support for missing features, and participate in the touch advances being made in open source world. To complete the stack, we’ve created an open source gesture recognition engine and defined a gesture API that provides a means for applications to obtain and use gesture events from the uTouch gesture engine."

to post comments

Multitouch support for Ubuntu 10.10

Posted Aug 16, 2010 16:34 UTC (Mon) by jspaleta (subscriber, #50639) [Link] (28 responses)

As a vocal critic of Canonical, its right and proper that I lift this effort up and say that I am tentatively hopeful that this effort will be the basis of a long term upstream solution.

http://voices.canonical.com/chase.douglas/2010/08/16/thou...

"So we have multitouch gestures through X, we’re done now right? Wrong. Our approach of embedding all of this can be termed as a “hack”, a suboptimal solution. We fully recognize this and include it in Ubuntu 10.10 only as a stop-gap measure. We will be reaching out to the X developer community to open a dialog over the potential inclusion of a more optimal solution in X, if that is desirable for everyone."

I very much like that Canonical recognizes that parts of this internal Canonical implementation are essentially throw away and a real solution on what the X extention that provides this functionality needs to be driven into X as part of a discussion with X.org developers. I look forward to watching that conversation happen in the X.org lists and codebase.

But one word of caution. If Canonical requires copyright assignment for geis and grail, the libraries that application developers will be asked to target, such a requirement will hamper development. If grail and geis are moving pieces which will need to adapt to upstream input it will be much more efficient if external contributors can effect changes to grail and geis as peer contributors and not have to assign copyright to Canonical.

It will be interesting to see if future versions of gtk fold in the functionality present in grail/geis once a stable X extension is available.

-jef

Multitouch support for Ubuntu 10.10

Posted Aug 16, 2010 19:58 UTC (Mon) by j1mc (subscriber, #56848) [Link] (18 responses)

Yes, it looks like contributions will be required to have the copyright assigned to Canonical.

Multitouch support for Ubuntu 10.10

Posted Aug 16, 2010 20:15 UTC (Mon) by jspaleta (subscriber, #50639) [Link] (17 responses)

Of course it does. But I want to make sure everyone is fully aware of that fact and what that implies for the future of these libraries. They are doomed to be integrated as upstream framework components because of the copyright control requirements. Even if these are worthwhile pieces of technology, by keeping copyright control now via an assignment policy, Canonical is knowingly making it much more likely that a fork or re-implementation that is incompatible with their libraries will end up as the upstream implementation for application developers to depend on.

This policy of copyright assignment works against the long term interests of both the operating system developers who do their contribution beneath these libraries and to application developers who do their contribution above these libraries.... all for the sake of preserving Canonical's ability to re-license these libraries under a proprietary license at some future date. The continued insistence of Canonical as a for-profit entity to require copyright assignment will just end up marginalizing their implementation and the ecosystem will either fork it or re-implement the functionality in a codebase that does not require assignment to a for-profit entity.

-jef

Multitouch support for Ubuntu 10.10

Posted Aug 16, 2010 21:02 UTC (Mon) by koverstreet (✭ supporter ✭, #4296) [Link] (2 responses)

So fork them. Everyone goes on about how the right to fork is so important to free software - but then the complain that forks represent some kind of failure... which becomes an excuse to not bother.

The freedom to fork doesn't mean a thing if you aren't willing to use it. Furthermore, tools in the past 5 years have made it a hell of a lot easier than it used to be.

There's been a lot of bitching and complaining about Canonical's copyright assignment, but as long as it's just talk they'll most likely ignore it. If people want it to change, fork it, use the fork upstream, and let _them_ decide whether it's worth it to them to maintain their own branch.

I think Canonical's copyright assignment policy is ridiculous and stupid, but it's not like they're being malicious; they're well within their rights to merge code based on whatever criteria they feel like. If nobody's going to challenge their ridiculousness, expecting them to change is about as ridiculous as they're being.

Multitouch support for Ubuntu 10.10

Posted Aug 16, 2010 21:06 UTC (Mon) by jspaleta (subscriber, #50639) [Link] (1 responses)

Who said they are being malicious? Canonical isn't Google.

And isn't publicly talking publicly about the policy part of challenging it?

-jef

Multitouch support for Ubuntu 10.10

Posted Aug 17, 2010 4:36 UTC (Tue) by drag (guest, #31333) [Link]

> And isn't publicly talking publicly about the policy part of challenging it?

A little bit. But Linux is a meritocracy, not democracy. So trying to sway opinion means very little in the long run. It's fun to have discussions and that sort of thing, but ultimately it's people that do the work that do the deciding.

Multitouch support for Ubuntu 10.10

Posted Aug 17, 2010 11:28 UTC (Tue) by hppnq (guest, #14462) [Link] (13 responses)

Netmail. Hula. Bongo. Novell.

Really, Jef, you don't have to frighten people with your imaginations and conjecture. You can just refer to the many other projects that require copyright assignments, for a host of reasons, and study what has become of them.

Apache. Mono. JBoss. GNU cat. The latter is doomed to remain integrated in the GNU utils, did you know that?

You were, of course, the first entry in my LWN killfile. This helps. But it only helps me: you are still able to use LWN as a platform to distribute all kinds of noise that do nothing to inform innocent readers, or, say, further the cause of Free Software.

So let me say this: it does not reflect well on the Fedora project and its sponsor Red Hat that you spew your interpretations and insinuations in public without properly stating that they are yours, and yours alone.

(They are, aren't they?!)

Multitouch support for Ubuntu 10.10

Posted Aug 17, 2010 17:40 UTC (Tue) by ovitters (guest, #27950) [Link] (12 responses)

I have understood from your message that you do not like Jef. Still, I'd appreciate if you could respond to him without using such a condescending tone.

Multitouch support for Ubuntu 10.10

Posted Aug 17, 2010 19:46 UTC (Tue) by hppnq (guest, #14462) [Link] (11 responses)

Insulting, perhaps. Condescending is telling other people what to do with their own code. Annoying is liberally applying double standards in the process. Questionable is repeating this ad nauseum.

But I really don't want to offend anyone, so if I rubbed you or anyone else the wrong way I hope you'll accept my apologies. For the record, I have nothing against Jef personally, and expect him to take this kind of criticism on the chin.

So how about that multitouch support, eh?

Multitouch support for Ubuntu 10.10

Posted Aug 18, 2010 14:35 UTC (Wed) by bronson (subscriber, #4806) [Link] (10 responses)

> I really don't want to offend anyone

You're kidding, right? From your loaded adjectives ("insinuations", "noise", etc), tone, and ad hominem style, it looks like you were making an effort to offend.

Multitouch support for Ubuntu 10.10

Posted Aug 18, 2010 23:40 UTC (Wed) by hppnq (guest, #14462) [Link] (9 responses)

You seem to have missed that I apologized to anyone who was disturbed by the tone or the content of the message that was directly aimed at Jef, even though I think I have every right to question in strong words the motive of someone who keeps spoiling perfectly technical items in a way that can be accurately described with very normal nouns like "insinuations" and "noise".

This is what Jef calls, interestingly, "fair game".

I will leave it at this.

Multitouch support for Ubuntu 10.10

Posted Aug 19, 2010 0:14 UTC (Thu) by jspaleta (subscriber, #50639) [Link] (8 responses)

Did you question my motives? I'm not sure you did. You did give a long list of projects which you claim have a copyright assignment policy similar in nature to Canonical's. I've been attempting to do my due diligence by reading over the contributor license agreements for those projects to confirm your assertions that they do have blanket copyright assignment clauses embedded in them. I'm having trouble finding them in some cases.

For example JBoss requires a contributor agreement for many of the projects under the JBoss umbrella. But I cannot find the clause in it which states that you are required to assign copyright to another entity.

For example JBoss Web requires in-part:
"You hereby grant to Red Hat, its successors, and assigns, the non-exclusive, transferable, irrevocable, perpetual, royalty-free right to use, modify, copy, sell, and distribute the Contributions under the terms of any version of the GNU General Public License, or any version of the GNU Lesser General Public License. Without limitation, this grant is made with respect to any copyright, patent, or other intellectual property or moral rights You may have in or to the Contributions."

That is a very narrowly worded licensing clause which does not give Red Hat the copyright control necessary to re-license a contribution under a proprietary license. Its important to understand the difference between this sort of language...language that prevents proprietary exploitation of contributed work...and a blanket copyright assignment which allows a central authority to create proprietary licensed version of contributed code. The option to re-license code under later version of the GPL and LGPL is a universe away from a blanket copyright assignment which would allow a central for-profit entity to re-license contributions however it desires. If Canonical used similar narrow language such as that found in JBoss Web's CLA there wouldn't be much of an issue.

-jef

Multitouch support for Ubuntu 10.10

Posted Aug 19, 2010 10:21 UTC (Thu) by hppnq (guest, #14462) [Link] (7 responses)

Did you question my motives? I'm not sure you did.

Oh, I did. Really. Still do.

You did give a long list of projects which you claim have a copyright assignment policy similar in nature to Canonical's.

The topic of copyright assignment has been chewed to death for years. I gave a list of projects that you could study to see how they employ copyright assignment. Your due diligence should have made it clear to you that all these policies are different, that all these projects and companies -- and there are many more -- have their own different reasons for requiring contributors to enter into a special agreement regarding the intellectual property of the code, and that the results differ wildly.

It should have made it clear that there is no particular reason to single out the multitouch libraries in the long list of actual, popular and possibly heavily integrated chunks of software that are perhaps also running on your system right now, to make people aware of the risks of copyright assignment in general, or this one in particular.

Am I correct in thinking that, with all those announcements of new releases of, say, Evolution, not once did you chime in to warn people about the dangers of assigning copyrights to Novell? Are you surprised that I find this strange and therefore question your motives for spreading this "helpful information", especially since Novell's track record can't have escaped you?

Now, JBoss, indeed, is a bit different story. The agreement you need to enter does not require you to transfer copyrights, it only wants to make sure the license remains a (L)GPL one. Oh, and all the files have to include a JBoss, Inc. copyright notice. Fair enough, no? Especially considering the JBoss, Inc business model. Red Hat bought JBoss, Inc and the community, no copyright assignment needed. It worked out quite successful for them, a prime example of how community resources can be used to maximize corporate profits.

Multitouch support for Ubuntu 10.10

Posted Aug 19, 2010 10:35 UTC (Thu) by rahulsundaram (subscriber, #21946) [Link] (4 responses)

"Have to" is very different from "should". It is merely a template as shown by the wording

"All files (including tests) should have a header like the following:"

It is merely sound legal practise to have a project name listed and a per file copyright notice giving you all the rights under GPL or LGPL as applicable.

Multitouch support for Ubuntu 10.10

Posted Aug 19, 2010 13:38 UTC (Thu) by hppnq (guest, #14462) [Link] (3 responses)

"Have to" is very different from "should". It is merely a template

Actually, I have checked quite a big handful of JBoss files and got a 100% score before I got bored: this template seems to be an awfully popular actual copyright notice.

It is merely sound legal practise to have a project name listed and a per file copyright notice giving you all the rights under GPL or LGPL as applicable.

To me it does not sound like a great idea, legally speaking, to put "Copyright EvilCorp" at the top of my file and relegate myself to @authors.

Multitouch support for Ubuntu 10.10

Posted Aug 19, 2010 13:57 UTC (Thu) by rahulsundaram (subscriber, #21946) [Link] (2 responses)

A popular notice isn't equivalent to a mandate and unless you sign a agreement to give your copyright to a corporation, authors retain their copyright. It is not possible for a mere template to change that.

Multitouch support for Ubuntu 10.10

Posted Aug 20, 2010 8:40 UTC (Fri) by hppnq (guest, #14462) [Link] (1 responses)

No one is refuting that. The JBoss story is about the "myth of Open Source".

Multitouch support for Ubuntu 10.10

Posted Aug 20, 2010 9:24 UTC (Fri) by rahulsundaram (subscriber, #21946) [Link]

The topic of the thread was copyright assignment and the article you reference doesn't seem related. Although only tangentially related, If you are talking about the differing business models, JBoss's original model created by Marc (who is no longer with Red Hat) was to employ all the major contributors and Red Hat doesn't have a mandate to do that anymore after the acquisition. So perhaps it is relevant as a historical note but I am not sure of the point you are trying to make.

Multitouch support for Ubuntu 10.10

Posted Aug 19, 2010 18:16 UTC (Thu) by jspaleta (subscriber, #50639) [Link] (1 responses)

Evolution is a very interesting example as a piece of history of a corporate entity coming to understand that blanket copyright assignment is a suboptimal way to build a vibrant contributor community around a codebase.

http://mail.gnome.org/archives/desktop-devel-list/2008-Ju...

A post made ironically enough on the same day I started posting on LWN. Difficult to argue that I missed a chance to lobby against Novell's copyright assignment policy on Evolution in commentary here when they decided to stop requiring about the same time I can to LWN. Maybe part of the reason they changed their policy was fear of my yet to be written scathing commentary of their policy.

Regardless of the reason, the history of evolution's change from a copyright assignment policy is a good lesson for Canonical to learn from for _all_ the codebases it requires copyright assignment for...a growing list of convenience libraries not limited to just this new multi-touch stack. It's a systemic problem in how Canonical approaches new development efforts.

-jef

Multitouch support for Ubuntu 10.10

Posted Aug 20, 2010 9:38 UTC (Fri) by hppnq (guest, #14462) [Link]

Wow, it feels like much longer.

Obviously, you're not going to humour me, so I took some time to investigate. How could you miss this opportunity for instance? That's a very enlightening comment.

I am done now!

Multitouch support for Ubuntu 10.10

Posted Aug 17, 2010 10:46 UTC (Tue) by bregma (guest, #69594) [Link] (8 responses)

The FSF requires copyright assignment for most of their projects (eg. GCC). In what way is that different?

The geis library is free software. Anyone is free to fork it. Anyone is free to provide a ground-up implementation of the same API without a single line of shared code and have client software use it at runtime without recompilation. Free as in RMS's vision of freedom.

Canonical requires copyright assignment for contributions to its implementation of geis and to grail. That means if a flaw in the GPL is discovered, the single copyright owner can alter the license terms to use the new, fixed GPL. This has historical precedence. Holding all copyrights in a project is the best way to ensure the software remain free, everywhere and forever.

And really, I don't see freedom as a bad thing.

Multitouch support for Ubuntu 10.10

Posted Aug 17, 2010 11:03 UTC (Tue) by tzafrir (subscriber, #11501) [Link] (5 responses)

The FSF guarantees that they will not hijack your code. Cannonical does not.

What happens if Oracle buys Cannonical?

Multitouch support for Ubuntu 10.10

Posted Aug 17, 2010 13:58 UTC (Tue) by hppnq (guest, #14462) [Link] (4 responses)

The FSF guarantees that they will not hijack your code. Cannonical does not.

The only sane reason to require copyright assignment, is that you can relicense the software covered by it. The FSF promises to not release the software under a non-free license, Canonical indicates that it may do exactly that. There is nothing mysterious about this.

There are several kinds of Free Software that permit this kind of behaviour, and this is certainly not universally recognized as hijacking. Although, believe it or not, some people get freaked out by the idea that their software might someday be released under GPLv4. As always, if dealing with a piece of software requires a certain amount of reading so you can enter an agreement, you should bloody well read it. Superstition ain't the way.

What happens if Oracle buys Cannonical?

Mark Shuttleworth is catapulted into space laughing hysterically and releases Ibinti upon his safe return?

Multitouch support for Ubuntu 10.10

Posted Aug 17, 2010 14:55 UTC (Tue) by Trelane (subscriber, #56877) [Link] (2 responses)

The only sane reason to require copyright assignment, is that you can relicense the software covered by it.

I disagree

Multitouch support for Ubuntu 10.10

Posted Aug 17, 2010 16:51 UTC (Tue) by hppnq (guest, #14462) [Link] (1 responses)

Over the years, there have been several examples of actual enforcement of the GPL where there were multiple copyright holders.

Multitouch support for Ubuntu 10.10

Posted Aug 17, 2010 22:30 UTC (Tue) by Trelane (subscriber, #56877) [Link]

Yes, but that doesn't make the FSF reason insane, as OP (likely unintentionally) suggested.

Multitouch support for Ubuntu 10.10

Posted Aug 17, 2010 15:34 UTC (Tue) by tzafrir (subscriber, #11501) [Link]

If and when Cannonical actually makes any legally-binding promise similar to that of the FSF, I will feel more comfortable about assigning code to it. Depending also on the content of that promise. Right now, they don't.

If Oracle buys Cannonical and decides to relicense future versions of the uTrace stack under the terms of the Oracle Orthodox Permissions System, it would be hijacking. Regardless of whatever stunts Mr Shuttleworth does later on.

Multitouch support for Ubuntu 10.10

Posted Aug 17, 2010 11:04 UTC (Tue) by rahulsundaram (subscriber, #21946) [Link]

FSF's copyright agreement is actually quite different.

http://ebb.org/bkuhn/blog/2010/02/01/copyright-not-all-eq...

LWN covered this at

http://lwn.net/Articles/359013/

Multitouch support for Ubuntu 10.10

Posted Aug 17, 2010 17:02 UTC (Tue) by jspaleta (subscriber, #50639) [Link]

From a risk management pov... generally speaking..assigning copyright to a not-for-profit represents lower risk of the code being re-licensed as proprietary code. Generally speaking... not-for-profits like the FSF have a mandate as part of their charter to do public good and are much less likely to be bought out by another controlling interest and change internal policy.

But more importantly, when reading any such assignment contract its vital that there be a guarantee for code access for all future versions of the code. The FSF in particular has strong guarantees in place which explicitly prohibit them from taking the code and re-license it in a proprietary manner which would later restrict access of the codebase to any and all contributors who have assigned copyright to the FSF.

This is exactly the sort of guarantee that is rarely found in the assignment contracts with for-profit entities...exactly because the for-profit entity _knows_ and _values_ the potential _profit_ to be made in a proprietary re-licensing and want to keep that option open to themselves..and only for themselves. What is thought of as enlightened self-interest to a not-for-profit with a charter to do public good can be very different than what is thought of as enlightened self-interest for a for-profit start-up desperately looking for a solid revenue stream.

The basic problem is that when any for-profit company stands itself up as a more than equal contributor and gives itself more flexibility to use the code than other contributors..its ultimately unhealthy for the software project. There is a history of this happening. Just look at Openoffice. Copyright assignment to a for-profit company actively dissuades external contributors from participating in that code development because of the less-than-equal nature of the relationship among contributors.

The one failing of the GPL is that in some ways it encourages these situations by giving for-profit-entities the ability to have their cake and eat it too via copyright assignment and dual licensing. If this library code were BSD there would be less of an issue here because such a license allows anyone to make a proprietary fork and run with it. Copyright assignment coupled with BSD code doesn't really have a downside because every contributor has equal rights to the BSD codebase (including making a proprietary product from it) even if they assign copyright to a central authority who later shutters the BSD version completely.

If the goal was wide distribution of this multitouch technology, then the libraries would be licensed with a much more liberal and business interest friendly license. But the choice of license and the requirement for copyright assignment is as much about control of the business opportunity as it is about technology innovation. History so far hasn't really all that kind to vendors who try to exert this sort of control. This multitouch stuff even if its very good technically will be marginalized and replaced with re-implementations unless Canonical lets go of its copyright assignment requirement for the convenience libraries that sit above X and below the application level. There's very little upside to controlling the copyright of the convenience library layer of technology as tightly as Canonical wants. Honestly assigning their copyright to the FSF for these libraries is actually overall a better strategy than holding the copyright themselves... if there really is concern about legal protection down the road. But in reality, these libraries may be better licensed BSD than LGPL for widest impact and uptake.

-jef


Copyright © 2010, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds