On 5/22/05, Brion Vibber <brion(a)pobox.com> wrote: Agreed. And here a derivative problem would rise: how we treat submissions from anonyomous (or unregistered users precisely). Curently more than half JA WP & EN WQ contributors are anons. On other projects, about 16-20%. I would make it sure we can safely ignore those anonymous submissions' right as "licence holder" [or somehow possible, because they can't or hardly be contacted, but I am not sure]... -- Aphaea(a)*.wikipedia.org email: Aphaia @ gmail (dot) com

Hallo, und thank you for your reply On 5/22/05, Timwi <timwi(a)gmx.net> wrote: Okay, so they hold their copyright until 50 years will pass from their submissions ... So rigidly it is impossible for us to switch licence -- or we should cast once all of anonymous submissions (because there is no established contact between them and us, and we are not sure talk page is enough for this purpose ...) -- Aphaea(a)*.wikipedia.org email: Aphaia @ gmail (dot) com

As for re-licensing Wikipedia, personally, if it was up to me, I would be "bold" and do it the following way. I know that this is legally less secure than what Mozilla did, but I think that given our situation, it is reasonable. Someone in this thread mentioned that there is a judge who gives legal advice to Wikimedia -- maybe if he/she could post to give their opinion on whether this is legally feasible. My idea is thus: * Announce publicly the intended change of license from GFDL to WPL (Wikimedia Public License), and invite contributors to complain if they are unhappy with their material being re-licensed. Delete and re-write the material that is complained about. * At some point (call it <date>), have all new articles licensed under WPL only. * For one year starting on <date>, display a clear message on every page created before <date> stating that "This material is GFDL. We are intending to re-license it as WPL on <date+1year>. If you are the copyright holder and do not wish for your contributions to be re-licensed, please contact us." * I imagine very few people will contact us, and we can state that people's silence is interpreted to mean they're OK with the re-licensing. * Announce all of this extremely loudly in public. Post to all sorts of message boards, have news sites report it, etc. We have enough publicity to claim that we have contacted all contributors via public means. At that point, I believe it is no longer our responsibility if someone didn't notice anything for a whole year. * One year after <date>, switch everything to WPL-only (or whatever you like). People may still complain after this, in which case we can still remove and re-write their work, but it is not really our responsibility that other people may have already re-used the work under WPL terms. Timwi

On another hand, we are all talking like if re-licensing should be done right now in the blink of an eye... when we have time and we can do it slowly. If we need something like a roadmap : 1. A good first step could be to enable dual-licensing as the default policy for new additions, 2. then to begin to authors of old-gfdl content to re-license their stuff, 3. and finally to replace gfdl-only content if necessary. That would allow a slow but smooth process for relicensing. With adapted tagging, it could be a very good solution. Jean-Baptiste Soufron Le 22 mai 05 à 09:01, Brion Vibber a écrit :

On 22/05/05, David Gerard &lt;fun(a)thingy.apana.org.au&gt; wrote: I don't think this is as good a fit as some people are suggesting, since it's based entirely around *opinion* - everything to do with the quality, appropriateness, etc of an article is quite rightly regarded as a subjective decision, to be taken collectively by the community. Whether or not a particular version of an article is distributable under a particular license (e.g. whether a given definition on a Wiktionary can be imported into a GFDL-incompatible database), on the other hand, is a completely *objective* question - the article either does, or does not, contain material contributed to by people who have not granted permission under licenses other than the GFDL. ---- Of course, this actually makes it far *easier* to do the tagging - the database "knows" who has contributed to an article (or revision), so if it was "told" which users had agreed to new licensing terms, it could determine (and update) the licensing status of every revision in the database. No human would need to think about it, just ask the computer! This leaves articles in one of three states: 1) The current revision is available under the new licensing terms (all contributors to date have agreed to adopt them). 2) The current revision incorporates material not available under the new licensing terms (contributed by users who have not agreed to adopt them). However, there is a revision before any such non-relicenses contributions were made; therefore, that revision is safe to distribute under the new terms. 3) The very first revision was contributed by a user who has not agreed to relicense; taking all subsequent revisions to be "derivative versions" of that original material, the entire article is unavailable under the new licensing scheme. For importing to Ultimate Wiktionary, this is all you need to know - at any point in time, you can act according to the state of an "article": If 1: import the current revision If 2: import the newest revision which is available under the new terms If 3: ignore the article, and create it from scratch within the new database ---- For relicensing Wikipedia, things become somewhat more complicated. Given the age of the project, and the general volume of edits and editors, there are likely to be a far higher proportion of articles in states 2 and 3, and by the nature of the content there would be a greater need to minimise the amount of effort "thrown away". One possibility is that the license-checking tool could be made aware of page reversions - so an article in state 2 could be reverted to a relicensable revision and then re-developed. Presumably, information added later by contributors who had agreed to the terms could legally be [probably manually] re-added, as long as the bits added by people who *hadn't* agreed *weren't* included, in any form. A key question would then be how to define the difference between a derivative work and a completely new one. How much does the article need to have changed from an older revision to claim that the newer version is not subject to the copyright of an older author? This would be particularly important for articles that were created a long time ago, and might well end up stuck in state 3 because the original author had long since ceased to be contactable - could the work put into vastly expanding Wikipedia's oldest articles be legally salvaged if everyone except their originators agreed to new terms? -- Rowan Collins BSc [IMSoP]

On 24/05/05, Brion Vibber &lt;brion(a)pobox.com&gt; wrote: Hmm, that is indeed a good and rather worrying point: content imported under the GFDL or similar and compatible licences probably *should* be creditted as such in some way which could easily be recognised - if not by computer, by a dedicated team of humans telling the computer. But material copied from other articles - merges, splits, not to mention translations - is generally treated very laxly, with hopefully a reference in the edit summary, but not always even that. Put it together with the thorny question of "when is a rewrite a rewrite", and it makes you wish for a meaningful "blame"/"who added this line" tool - though I tend to agree with the opinion that this would be an order of magnitude harder for the free text of encyclopedia articles than it is for source code. Although, it has to be noted that those IBM researchers managed to get meaningful data in their "history flow" system... My point being, that if relicensing of this sort *were* ever attempted on Wikipedia, it would be impossible to do it properly without some such tool to trace where every bit of the text came from, and to check (manually, since edit summaries aren't likely to be reliably machine-readable) that the crucial edits were not only made by new-license-agreeable editors, but came from sources which were themselves new-license-compatible. -- Rowan Collins BSc [IMSoP]

On 26/05/05, David Gerard &lt;fun(a)thingy.apana.org.au&gt; wrote: Yes, but in free natural language text like a Wikipedia article, there isn't a meaningful definition of a "line" like with source code - you've got to either look at a paragraph (probably too big), or maybe a sentence (which requires somewhat more complex parsing - how many sentences are there in "e.g. What's the magic no.? 1.23!"). Plus, the database needed to carry that granularity of blame between articles in the whole of Wikipedia would surely be humongous - presumably involving some index of every line of every revision of every article in the entire encyclopedia, ready for comparison. But like I say, I don't know how the IBM folks did it, nor even how "normal" source-analysis tools work, so maybe it is possible after all. -- Rowan Collins BSc [IMSoP]