I was just made aware of this thread, and I realise that potentially a legal issue is discussed on wikitech. I would like the opinion of our lawyers on this specific point.

So, tel me if I understand well, to comply with the gfdl the best we can (and we already know it is problematic), what you suggest is to list first the real name contributors, followed by pseudonymes, then by ips. Of course, the number of names is limited. We can expect that on many articles, the number of names will be over 50 or more.

I understood the gfdl "normal" requirement is to list the 5 main contributors. We probably know that we can define who the 5 main contributors are. Indeed, unless the number of contributors is below 5, there is no way to report with honesty the legal requirements.

This said, if we can't report reality, why would we report a group of contributors more than another ? If a pseudonyme wrote 95% of an article, and 5% officially real names corrected typos, is that really correct to indicate these 5 real names and not the pseudonyme ?

I would say it is not. Legally, that is incorrect.

From a community view point, that is setting a case

which I am not sure is really positive. It think that it would be more correct to make random choice among pseudo or real names, or to choose among the last ones.

I will forward this to wikipedia-l and foundation-l, since I believe this is more than a technical issue.

Evan Prodromou a �crit:

So, I'd like to add a little block of attribution

data to each page

(optional, per-installation; I'm guessing Wikipedia

wouldn't use

this). Something along the lines of:

   This article last edited on April 21, 2004 by

Evan Prodromou.

   Based on work by Alice Notaperson, Bob

Alsonotaperson, users

   Crankshaft, Deckchair and Eggplant, and

anonymous editors.

For each (distinct) person who's listed in the old

table, it'd show

their real name if it's set, or their user name if

not. All anonymous

edits would be lumped under "anonymous editors".

Contributors would be

listed with real-named folks first, then pseudo'd

folks, then

anonymous. There's no particular reason for that; it

could be any

other way (although I don't see a big point making

it configurable).

The goal here is to make it easy for redistributors

to comply with

license provisions that require author attribution

(such as some

Creative Commons licenses), without having to dig

through a whole

bunch of history pages.

Anyhoo, the Metadata.php code already does most of

this logic, albeit

for output in RDF format. I'd like to take that

stuff and put it in

the Article class, in a method like

"getContributors". The method

could then be used both from the attribution code

and from the RDF

metadata code.

getContributors would return an array of arrays,

each of which would

contain:

    0. User ID
    1. User account name
    2. User real name, if set

Another option would be to create User objects for

each entry in the

returned array, but a) I don't think that most of

the User object

fields (email, preferences) are needed, and b) I'd

be worried about

slingin' around incomplete User objects. So, I think

the arrays are

the best bet.

Does returning an array of arrays seem insane? Would

it be wrong to

add this method to Article? If so, where else would

it go?

~ESP

__________________________________ Do you Yahoo!? Win a $20,000 Career Makeover at Yahoo! HotJobs http://hotjobs.sweepstakes.yahoo.com/careermakeover