Informatics Educational Institutions & Programs

Since September 2006, the Free Software Foundation has been asking for comments on the GNU Simpler Free Documentation License. This is the license we will most likely migrate to with the next update of the GFDL. This being the case it is important that we make sure that the license is as well written as possible. Existing suggestions may be found at m:GFDL suggestions. Further suggestions are requested below.

At the present time I would suggest out priorities are getting rid of the remaining invariant sections (disclaimers and copyright notices) and simplifying the attribution requirements.Geni 02:48, 5 October 2007 (UTC)[reply]

Preamble

The GNU GPL [1] has a great preamble. If you've never read it, do so now; it's only about 500 words. The preamble explains the purpose of the license without diving into the legal jargon. The revised GFDL/GSFDL deserves a good preamble. The existing GFDL preamble is too focused on only one class of use, and it's just not as good.

I think a quality preamble does a lot to improve the human-readability of the license, since most of the time people don't care about the details. The Creative Commons licenses have a 'human readable' version of their terms, but I think that's a bad idea, because people often confuse them for the actual license requirements. (For example, the cc-by human-readable text says that you get attribution "in the manner you specify", causing people to claim that they can specify they be attributed via skywriting or whatever. ;) ). — Preceding unsigned comment added by Gmaxwell (talk • contribs)

Title Page / Author List

This, like the GFDL, envisions the existence of a "title page" providing copyright, license, and authorship info. Wikipedia has never really had this structure. In particular, though the history section probably meets the spirit of the GFDL, it is unclear whether having a seperate history page with a chronological contributor list really meets the specific structure described in the G(S)FDL. It would be nice if this the license were tweaked to be more unambiguously consistent with Wikipedia's version of giving credit. 69.226.73.169 04:43, 5 October 2007 (UTC)[reply]

Name of work

"4A. Use a title distinct from that of the Work, and from those of previous versions of the Work as listed in the History section". Generally all versions of a Wikipedia article are identified by the same title and hence do not comply with this. 69.226.73.169 04:43, 5 October 2007 (UTC)[reply]

Each revision is assigned a specific identification number (the last one prior to this post for this page is 162398612 for instance)... I don't know if that counts, though. --Jeffrey O. Gustafson - Shazaam! - <*> 05:43, 5 October 2007 (UTC)[reply]

Sections = articles

It would be nice for the discussion of "sections" to have "(or articles)" or the like so that it would be obvious that Wikipedia, like any other encyclopedia, as a whole is one work. This would mean that we don't have to artificially mess with histories to preserve attribution when, e.g., articles are split or merged. 1of3 21:18, 5 October 2007 (UTC)[reply]

That's exactly backward. Wikipedia should be doing more to make authorship of its various articles and sections transparent, not asking others to change the concept of authorship so we don't have to. Even encyclopedias will list the author(s) of their articles in the article itself. (For Brittanica, these are at the end of each article, though in their case it is generally only one or a few names.) Wikipedia needs to have better tools for tracking authorship in the event of such splits and merges. 76.240.228.205 05:22, 7 October 2007 (UTC)[reply]

Excerpts

The new rules for excerpts are fairly useless as they limit excerpts to 20,000 characters or 12 printed pages, which probably excludes virtually all of our best articles (FA-class). This needs to be changed to either 200,000 characters or 20,000 words, as the entire point of this section, as I understand it, was to allow reuse of individual Wikipedia articles without burdening them with including the entire text of the GFDL. For printed pages, I would suggest changing it to 50 pages, which should accomodate virtually all individual articles (see World War II for example). Kaldari 23:47, 15 October 2007 (UTC)[reply]

20,000 words is greater than some plays. The problem is that some wikipedia articles are so long they can exceed short books in length. It is not imposible that they exceed in length any manual ever released under the GFDL.Geni 00:28, 16 October 2007 (UTC)[reply]

Why is 20,000 words a problem, but not 20,000 characters? If I wanted to print out the article on Armenian Genocide (80,000 chars; 12,000 words; 25 pages) and give it out to a 10th grade history class, I shouldn't have to include 10 pages of lincense text. A license URL should be sufficient for a single article. (Unfortunately, we can't say "single article" since such a thing is not defined in the GFDL.) Otherwise we are defeating the whole purpose of Wikipedia—providing information that is easy to use and disseminate across the globe. What is the point of even including the excerpts section in the license if it excludes all of Wikipedia's most useful and well-written content? If 20,000 words is excessive, what would be more acceptable? 10,000 words? 100,000 characters? 20,000 characters is just rediculously limiting, IMO. Have we really waited 5 years for such a meager concession? Kaldari 16:13, 16 October 2007 (UTC)[reply]

Hey I didn't say I objected to 20K words just going through the problems it would raise for other people.Geni 17:04, 16 October 2007 (UTC)[reply]

If making the GFDL work for Wikipedia is incompatible with other uses of the license, we need to abandon the GFDL (I know, easier said than done). In the meantime, we should be lobbying for whatever changes to the license are necessary to bring Wikipedia in line with the goals of the Wikimedia Foundation and the community. Token modifications are not going to help us. We need to lay out on the table the kind of license we desire and let the FSF balance that with whatever needs are expressed by other groups. Personally, I think we should be requesting a limit of no less than 100,000 characters for excerpts, preferably 200,000. Kaldari 19:31, 17 October 2007 (UTC)[reply]

The whole idea of the GSFDL as distinct from the GFDL is that the S version will be simpler to comply with. If this license is only suitable for use with wikis it will still have very widespread application - there are a lot of wikis and there will be more in future. One way to define 'excerpt' is as a percentage of the whole so an excerpt of a short work cannot be the entire work. I believe US copyright law has something like that to help define what is fair use. For wikipedia this has problems. I think that each wiki article is a separate work, with it's own title and history and discussion. I believe that anyone should be able to take any part or all of Wikipedia and reuse it provided they provided they offer their version under the same license (a link to the license text should be acceptable) and give attribution and that these requirements should be simple enough that they can be easily applied to even the shortest excerpts. Reuse of short excerpts without relicensing under GSFDL would then be limited to what copyright fair use allows in your country and need not be defined in the GSFDL. Filceolaire 09:40, 27 October 2007 (UTC)[reply]

Yes, I like your idea even better: "anyone should be able to take any part or all of Wikipedia and reuse it provided they offer their version under the same license (a link to the license text should be acceptable) and give attribution". Personally, I think providing a URL to the license text should always be adequate rather than requiring the reproduction of the license in full. I really don't see what good reproducing the entire license could possibly serve anyway. It's not like the license text is hard to find. Kaldari 18:16, 5 November 2007 (UTC)[reply]

Attribution

I think attribution should be satisfied by either

• Option 1. A general attribution to "Wikipedia editors" with a link back to the specific articles used (this then links to the history page, with user names, which then links to User pages with user attribution details). (Note that Wikipedia does not own the copyright to articles. Each editor owns the copyright to her own contributions)
• Option 2. A copy of the history page, as it was when the article was copied, attached to each derivative work. If offline then the relevant user pages should also be attached.

Users should have the option to chose either one. Wikipedians create derivative articles all the time. For translations into other languages option 1 is the appropriate route.

When an article is split I believe option 2 would be appropriate - i.e. both new articles inherit the old history.

When articles are merged option 2 would seem to be the way as well - with the new history page referencing the old (in the comment?). For distributing an article in class option 1 is appropriate,provided you spell out the url for each page used.

If you take a bunch of (GSFDL licensed) illustrations from commons and make a collage with a soundtrack of you reading random discussion pages and sell it to the Tate for a million bucks then the plaque they to attach to it will have to list all the contributors,not just you, and anyone can buy a copy in the gallery shop and make more copies.

For reuse as a script for a documentary film option 1 is the way to go with every article used listed in the credits (including the history page in the credits isn't required but will get you extra geek points).

If you scrape all the infoboxes and use these to create a catalog of everything then every datum needs to have a field linking to the relevant page or to a copy of it or else you need to do some data mining to identify the users who contributed that data and credit them individually.

If you use wikipedia to educate an artificial intelligence to act as your personal assistant, using the discussion pages as practical example of moral and ethical methods of resolving disputes (call it Jiminy Wikit), then Option 1 is satisfied if that AI mentions specific examples informing their advice, where relevant, but need not do so if their advice is based on a a broad trend or impression (no more than a human intelligence would be expected to). Note that, to meet the reuse under GSFDL criterion the knowledge derived by the AI from wikipedia's facts must be released for reuse by other AI researchers. (yes I do believe the first AI is more likely to be ruled by the Five pillars of Wikipedia than by the Three Laws of Robotics). Filceolaire 09:05, 27 October 2007 (UTC)[reply]

Revision process

How is the FSF actually working with and communicating with the Wikipedia community regarding the GFDL? I imagine the FSF is currently working on a second draft of the GFDL2 (and GSFDL2) now that they have GPL3 out of the way, but the communication channels seem to be one-way for the most part, i.e. I haven't seen any actual discussions between the Wikipedia community and the FSF about the licenses, ever. Maybe I'm too spoiled by Wikipedia's open and transparent nature, or maybe I'm just not aware of what's actually going on, but I find it surprising that there has been so little discussion with the FSF a year into the revision process (or if there has been, that that discussion has not been shared with the community). Kaldari 16:42, 16 October 2007 (UTC)[reply]

Last we heard a FSF board memeber was going to try and get a statement that would have been augustr the 24th.Geni 17:06, 16 October 2007 (UTC)[reply]

I've requested updates from the FSF on the status of the draft several times over the past year with no luck. The only response I've gotten is that I have to ask RMS about it, as if I had his address in my rolodex :P Kaldari 19:29, 25 October 2007 (UTC)[reply]

Wikipedia has contacts who will do.Geni 19:35, 26 October 2007 (UTC)[reply]

Why so cloak and dagger? I feel like I'm trying to ask the FBI about where to find nuclear weapons or something. When you say "Wikipedia has contacts who will do", who exactly is "Wikipedia", who are our "contacts", and what exactly are they going to do? Kaldari 20:47, 26 October 2007 (UTC)[reply]

Realisticaly sod all. Last comment I've seen is this.Geni 23:25, 26 October 2007 (UTC)[reply]

I sent emails to both Richard Stallman and Mako last month, but haven't received any response from either. Has anyone else had any luck? Kaldari 18:18, 5 November 2007 (UTC)[reply]

Changing of license

Is it possible for the license of all work on Wikipedia to change without permission of all contributors - while I am probably incorrect - it is concerning that Wikipedia may become split, with old content on one license and new content on another. This is far from desirable. MinuteElectron 22:19, 24 October 2007 (UTC)[reply]

Section 8a of the new version of the GFDL (which will presumably be finalized at the same time as the new GSFDL) states: "If the Work has no Cover Texts and no Invariant Sections then you may relicense the Work under the GNU Simpler Free Documentation License." Thus all Wikipedia content would be relicensed under the GSFDL. Kaldari 22:53, 24 October 2007 (UTC)[reply]

Compatibility with Creative Commons

Is there any way GSFDL can be made compatible with some Creative Commons license? Right now, we are in the ridiculous position where supposedly free content often requires author approval for redistribution because of the large variety of incompatible licenses. For example, I (usually) can't port content from Wikipedia to WikiTravel, because the latter uses the CC by-sa-1.0 license. The whole point of free licensing was to avoid this hassle. --Zvika 06:53, 25 October 2007 (UTC)[reply]

Given that CC licenses are not exactly stable and the philisophical differences between CC and the FSF probably not at this time.Geni 19:34, 26 October 2007 (UTC)[reply]

Look, I have no stakes in either FSF or CC. As far as I'm concerned all my contributions could be placed in the public domain. But I'm telling you, you are shooting yourself in the leg if you ignore other copyleft licenses. When I first heard of GFDL, I thought the idea was that I could do anything I wanted to with the content. Little did I know how difficult it would be to transfer content between two sites, both of which include only so-called "free" content. IANAL, but I really don't see what the big problem is if GSFDL would authorize redistribution under a particular version of a CC license. --Zvika 07:44, 27 October 2007 (UTC)[reply]

CC tend to focus on authors FSF tend to focus on reusers. This results in a degree of mutal distrust.Geni 11:52, 27 October 2007 (UTC)[reply]

Great news! It looks like we may be moving to a CC license after all! http://wikimediafoundation.org/wiki/Resolution:License_update. Kaldari 17:02, 3 December 2007 (UTC)[reply]

That really is great news! I hope this works out. --Zvika 18:11, 3 December 2007 (UTC)[reply]

The GSFDL adds a new section called "Free Manuals Are Essential" (which is not in the GFDL). It's basically nothing but 4 paragraphs of propaganda (specifically about manuals, not documents) that add nothing to the license. This philosophical waxing has no place in a license, much less a "simpler" license. Kaldari 20:42, 9 November 2007 (UTC)[reply]

I think this is widely agreed upon.Genisock2 19:52, 14 November 2007 (UTC)[reply]