Time problem [LWN.net]

Time problem

Posted Dec 16, 2012 18:18 UTC (Sun) by rfontana (subscriber, #52677)
In reply to: Time problem by man_ls
Parent article: Fontana: What open source licensing could learn from Creative Commons (Opensource.com)

Oh, that reminds me of a point I've thought about but neglected to mention in that article. One of the problems (I increasingly believe) with free software licenses is that they are updated too infrequently. (One may counterargue that this is one of their great strengths, at least for the more widely-used licenses; maybe both points are true.) The CC license suite upgrade cadence, at least as a matter of practice, is quite different from that of comparable establishment-institution FLOSS licenses

to post comments

Time problem, or perhaps not

Posted Dec 16, 2012 22:04 UTC (Sun) by man_ls (guest, #15091) [Link]

Given that relicensing is such a difficult feat in Free software, it is no wonder that people are not keen to adopt new licenses for their existing projects. True copyleft licenses are not backwards-compatible, and giving a "version x or later" carte blanche to organizations is not palatable to many developers.

Actually, the proliferation of Free software licenses is currently quite a problem, bearable only because 80+% of projects use the well known GPL(v2 or v3) and most of the rest use MIT or (2/3)-clause-BSD; corporate licenses account for a small percent of all projects. Now imagine if there were many updates with small but significant differences. (If the differences are not significant, then why bother?)

For many people even the GPLv2 works well enough so not everyone is willing to upgrade. Now imagine if there was a new set of licenses (GPL, LGPL, GFDL, and Affero) every year: a new schism for every ideological detail which is embedded in them. Not nice.

But the main reason for the slow upgrade cadence is the lack of perception of existing problems. With the GPL it has taken many years to update it, and apparently both v2 and v3 work well for their users. Why keep updating it?

The real cadence in Free software is not measured by its licenses but by its projects and communities. 8 years ago SourceForge was all the rage when now it's a shade of itself; people used Subversion and git didn't even exist; the most used distro was Mandrake (I believe) and the first Ubuntu version had just been released, while Android (a newcomer in 2004) is now used on hundreds of millions of terminals worldwide.

Meanwhile, in the world of CC licenses they may be updated frequently and are used by more and more people, but there are not many news of interest that I can think of. 8 years ago they were used by some people and now they have many more users. Yawn.

Time problem

Posted Dec 17, 2012 23:16 UTC (Mon) by louie (guest, #3285) [Link] (7 responses)

The CC license suite upgrade cadence, at least as a matter of practice, is quite different from that of comparable establishment-institution FLOSS licenses.

FWIW, I am pretty sure they see this as a bug and would like to move to a more GPL/MPL-like cadence in the future. (This will lock them into bad ideas, like the DRM clause, but the evidence suggests they are already locked into those.)

Time problem

Posted Dec 18, 2012 4:34 UTC (Tue) by mlinksva (guest, #38268) [Link] (6 responses)

I don't see a cadence. Fontana should describe that.

The months between versions so far have been 1.0->2.0: 17 months, 2.0->2.5: 12 months, 2.5->3.0: 20 months. 3.0->4.0 will probably be 60 or so months, ie 4.0 will optimistically be released early next year. An explicit goal of 4.0 is longevity; http://wiki.creativecommons.org/4.0#Goals_and_objectives doesn't say how long, but I hope for 10+ years. Each version has been prompted for a particular set of reasons.

I don't think more than a tiny improvement re DRM will be achieved, but there is still time to give feedback on this and all else, see http://wiki.creativecommons.org/4.0#Items_for_discussion

Time problem

Posted Dec 18, 2012 4:38 UTC (Tue) by louie (guest, #3285) [Link] (5 responses)

The cadence appears to be "double-digit months", which is... not the case for MPL, GPL, or Apache. :)

My understanding is that the DRM discussion is closed, certainly, I don't see any new arguments that could be provided, proved, or otherwise prove persuasive.

Time problem

Posted Dec 18, 2012 5:35 UTC (Tue) by rfontana (subscriber, #52677) [Link] (4 responses)

I guess that's all I really meant. The CC licenses seem to be updated more frequently than comparable (institution-associated?) free/open source software licenses.

Why not a rolling release license? :-)

Time problem

Posted Dec 18, 2012 7:58 UTC (Tue) by man_ls (guest, #15091) [Link] (3 responses)

You still have not answered: what is the advantage of having licenses that are upgraded fast? Or of rolling back changes? Are there so many pressing problems in the licensing world that they need to be addressed this fast?

I understand the proliferation of licenses (and even worse a rolling licenses) would create a myriad of legal interesting problems and a business opportunity for lawyers. Imagine that: "Judge, from 2012-12-18 12:47 to 2012-12-19 17:38 the license gave the right to distribute modified versions without attaching source code, and that is what my client did".

Disclaimer: IANAL, and most of my knowledge of US law comes from watching The Good Wife episodes.

Time problem

Posted Dec 18, 2012 14:29 UTC (Tue) by rfontana (subscriber, #52677) [Link] (2 responses)

You still have not answered: what is the advantage of having licenses that are upgraded fast? Or of rolling back changes? Are there so many pressing problems in the licensing world that they need to be addressed this fast?

"Pressing"? Probably not, though I wonder whether some minor fixes could have significant positive impacts on uptake of specific licenses.

In the proprietary commercial licensing world one doesn't typically see a form license in wide use kept entirely unmodified for, say, 15 years (indeed, licenses don't really have independent existence in a sense that is perceptible to the public; the release of a new product or new version of a product will be accompanied by at least some minor changes to prior versions). This is not necessarily because there are pressing problems to fix, but there are always things worth fixing if you have the ability to do so. As for negotiated license agreements in use for many years, these often see multiple amendments.

I suppose that isn't a satisfactory answer to your question about advantage. I feel that if you see a flaw in a license (I can tell you that in all widely-used licenses there are flaws, though they may not be 'pressing') it's a shame that you "can't" (in some cases this means "won't") fix it (other than, in some cases, through some cumbersome technique of issuing exceptions or clarifications). The idea that we should have to live with what turns out to be a mistake for 15 years, because it's so important to have a degree of legal simplification and stability that has no counterpart outside the FLOSS world, and because it's so important to standardize on a small set of licenses increasingly controlled by institutions, just seems questionable to me. That probably made more sense when "open source" was perceived as being younger and more novel. We are past adolescence.

I understand the proliferation of licenses (and even worse a rolling licenses) would create a myriad of legal interesting problems and a business opportunity for lawyers.

I don't actually think it would mean a net increase in interesting legal problems. The same set of interesting problems exist today because of license stability (and, increasingly, license standardization). Whatever business opportunities "open source" provides for a lawyer in private practice will, I expect, remain constant no matter what happens.

most of my knowledge of US law comes from watching The Good Wife episodes.

That's your problem right there. The Good Wife provides a highly unrealistic portrayal of lawyers, and particularly of lawyers in Chicago.

Time problem

Posted Dec 18, 2012 17:06 UTC (Tue) by man_ls (guest, #15091) [Link]

Thanks for your thoughtful and explanatory comment.

Time problem

Posted Dec 18, 2012 20:00 UTC (Tue) by apoelstra (subscriber, #75205) [Link]

> I suppose that isn't a satisfactory answer to your question about advantage. I feel that if you see a flaw in a license (I can tell you that in all widely-used licenses there are flaws, though they may not be 'pressing') it's a shame that you "can't" (in some cases this means "won't") fix it (other than, in some cases, through some cumbersome technique of issuing exceptions or clarifications). The idea that we should have to live with what turns out to be a mistake for 15 years, because it's so important to have a degree of legal simplification and stability that has no counterpart outside the FLOSS world, and because it's so important to standardize on a small set of licenses increasingly controlled by institutions, just seems questionable to me. That probably made more sense when "open source" was perceived as being younger and more novel. We are past adolescence.

I would argue that from a developer's perspective, the simplicity of having very few (and well-differentiated) licenses is very useful. It means that if you're just looking to link a library or lift code from somewhere, you can read just the title of a project's license to know if you need to keep looking.

And you also know roughly what you're allowed and not allowed to do, without being able to understand the legalese completely.

Having few organizations behind these licenses means that for the weird cases, you know who to ask for clarification.