The Cunctator wrote:
I strongly think we're better off formalizing a policy in which particular author attribution is not required. That is, by contributing to Wikipedia, you agree to be attributed as one of the "Wikipedia Contributors" or somesuch.
Jimmy responds:
Well that certainly much more closely matches our social custom, in which articles aren't "owned" by anyone, and we value all sorts of contributions to the project without specifically privileging 'authorship'.
But, how can we reconcile your suggestion with the FDL?
Not to be a wag, but how can we reconcile anonymous contributions of any stripe--including ones that aren't logged in--with the FDL? Pick an AOL IP address and try to assign it to one author, without AOL's help. Impossible. It doesn't even trace to one *computer* without AOL jumping through hoops, much less to one author at that one computer. (If this weren't the case, we wouldn't have the problem of recurring vandals).
Now, if the {irony} brilliant, uninvasive {/irony} idea of prepending IPs with phone numbers ever comes to fruition, maybe we would be closer to compliance with what the FDL seems to require. As it is, I think we're probably not, but also unable to do much about it without severely changing what wikipedia is.
kq
__________________________________ Do you Yahoo!? SBC Yahoo! DSL - Now only $29.95 per month! http://sbc.yahoo.com
koyaanis qatsi wrote:
Not to be a wag, but how can we reconcile anonymous contributions of any stripe--including ones that aren't logged in--with the FDL?
Those are pseudonyms. Nothing about the FDL requires authors to give their 'real names'. However an author chooses to identify themselves is good enough.
Pick an AOL IP address and try to assign it to one author, without AOL's help. Impossible.
Sure, and lots of people have the same names, too.
The GNU FDL only requires redistributors to pass along at least 5 names of prior contributors, not that the prior contributors be hunted down and investigated for their 'real' names.
----
I think this is worth stressing. Suppose someone came up to you and handed you a bit of paper with an essay on it. "Hi, my name is Bob, and you can create derivative works from this paper if you like, but you have to attribute the original to me. Bye bye!"
Are you ethically required to research whether or not Bob is his real name? Or, since he said he was 'Bob', isn't that sufficient, morally, ethically, and legally?
I think this is a non-issue, really.
--Jimbo
On Thu, 10 Jul 2003, Jimmy Wales wrote:
koyaanis qatsi wrote:
Not to be a wag, but how can we reconcile anonymous contributions of any stripe--including ones that aren't logged in--with the FDL?
Those are pseudonyms. Nothing about the FDL requires authors to give their 'real names'. However an author chooses to identify themselves is good enough.
An author does not choose their IP address. In the vast majority of cases, they won't even know their IP address, or even what an IP address is! You cannot claim that an author is "choosing" to use a name if they don't even know what that name is.
Pick an AOL IP address and try to assign it to one author, without AOL's help. Impossible.
Sure, and lots of people have the same names, too.
It is common in ordinary printed works for authors with names shared by other authors to be disambiguated in some way, for example by the use of middle initials. It is also common to have a paragraph of text about the author, saying who they are, possibly even explicitly telling people not to confuse them with another identically named person. Reasonable lengths are usually gone to to ensure that different authors are distinguished from each other.
If, as you seem to be asserting, the attribution requirements of the GFDL allow firstly for an author's "name" to be assigned by an external party, without the author's knowledge or consent, and secondly for no effort to be gone to to ensure that the "name" thus assigned corresponds to a single individual, then it follows that citing "the author" as the author of any piece of work is perfectly acceptable. If this sort of thing was really what the authors of the GFDL intended to be allowed, then they wouldn't have even included a requirement for authors to be listed. Interpreted as you seem to be interpreting it, the requirement would be entirely redundant. Clearly this is not what the authors of the GFDL intended.
I think any reasonable interpretation of the idea of listing authors would have the following two properties:
(1) The name of the author, even if it is a pseudonym, should be agreed by that author.
(2) Reasonable lengths should be gone to to make the name as near to unique as is feasible. (In a database, there is no reason for this not be become just plain "unique".)
The GNU FDL only requires redistributors to pass along at least 5 names of prior contributors, not that the prior contributors be hunted down and investigated for their 'real' names.
Even if it is agreed that pseudonyms are acceptable, it is still true that the current system doesn't even require *them*.
Perhaps the edit page should have a notice saying, "This edit will be attributed to you under the pseudonym [some unique identifier here]. If you do not wish to use that pseudonym, you may wish to register a user name or, if you already have one, log in."
That way, by clicking on "Save page", the user really would be agreeing to use this as their pseudonym. (It would also alert registered users who had been automatically logged out to the fact that this had happened. This has happened to me before, and I haven't noticed.)
Then you really could claim that the user has chosen to use this as a pseudonym, and different pseudonyms would correspond to different people.
Oliver
+-------------------------------------------+ | Oliver Pereira | | Dept. of Electronics and Computer Science | | University of Southampton | | omp199@ecs.soton.ac.uk | +-------------------------------------------+
Oliver Pereira wrote:
An author does not choose their IP address. In the vast majority of cases, they won't even know their IP address, or even what an IP address is! You cannot claim that an author is "choosing" to use a name if they don't even know what that name is.
If that's true, then we're free to identify them as 'anonymous', right? Or to identify them by the number that we have for them.
There's *nothing* that can give us a moral duty to try to found out the real name of a contributor, if they don't give it to us. If they don't give us a name, we are perfectly free to identify them by whatever we *do* have.
If, as you seem to be asserting, the attribution requirements of the GFDL allow firstly for an author's "name" to be assigned by an external party, without the author's knowledge or consent, and secondly for no effort to be gone to to ensure that the "name" thus assigned corresponds to a single individual, then it follows that citing "the author" as the author of any piece of work is perfectly acceptable.
No, that's not what I'm claiming at all, and the absurd conclusion that you cite does not follow from what I'm claiming.
What I'm asserting is that the GFDL requires us to identify up to 5 authors. This requirement can only mean that we are required to identify them using whatever information they gave us. If they didn't give us information, then we should identify them as best we can.
(1) The name of the author, even if it is a pseudonym, should be agreed by that author.
And that's exactly what I'm arguing for. If someone contributes anonymously, by not logging in, then we identify them as best we can.
(2) Reasonable lengths should be gone to to make the name as near to unique as is feasible. (In a database, there is no reason for this not be become just plain "unique".)
For example, we can use what little identifying traces they did leave us, i.e. their ip number.
Even if it is agreed that pseudonyms are acceptable, it is still true that the current system doesn't even require *them*.
No, it does require them. Every contribution is identified in a particular way.
This is a pretty silly argument, I think. What are we concerned about, exactly? Do you think someone could plausibly contribute an anonymous edit and then get mad at us later because we didn't do the impossible, i.e. we didn't hunt them down and find out their real name or get permission to identify them as best we could?
That's absurd, I think. We don't need to worry about it, and the FDL in no way requires such nonsense.
--Jimbo
Oliver Pereira wrote in part:
An author does not choose their IP address. In the vast majority of cases, they won't even know their IP address, or even what an IP address is! You cannot claim that an author is "choosing" to use a name if they don't even know what that name is.
True, the author licensed their work to the world completely anonymously, so I have no obligation to list their authorship except otherwise. That said, any author familiar with the practices of Wikipedia has a reasonably expectation that they will be identified by their IP address, so I believe that I really ought to identify them thus all the same.
It is common in ordinary printed works for authors with names shared by other authors to be disambiguated in some way, for example by the use of middle initials. It is also common to have a paragraph of text about the author, saying who they are, possibly even explicitly telling people not to confuse them with another identically named person. Reasonable lengths are usually gone to to ensure that different authors are distinguished from each other.
It's up to the author to do this, not the deriver. But again, authors familiar with Wikipedia practice can reasonably expect a link to thier user page, where they might place information of just this sort. That's why I think that derivers should link to Wikipedia rather than just copying down the author list.
-- Toby
koyaanis qatsi wrote:
Not to be a wag, but how can we reconcile anonymous contributions of any stripe--including ones that aren't logged in--with the FDL? Pick an AOL IP address and try to assign it to one author, without AOL's help. Impossible. It doesn't even trace to one *computer* without AOL jumping through hoops, much less to one author at that one computer. (If this weren't the case, we wouldn't have the problem of recurring vandals).
Well, I think the FDL only requires that you attribute the authors the same way they attributed themselves. My concern here then wasn't that it'd require too much attribution, but too little: if right now Wikipedia's authors aren't attributed in a list of authors, the authors of derived works could reasonably argue that they don't have to make such an attribution either. I think most of us would be happy with simply an attribution to "Wikipedia," but legally Wikipedia is not one of the authors, since the license explicitly says that you don't transfer copyright to Wikipedia when you post. I think legally Wikipedia is just another user of your GFDL text -- exactly the same as anyone else who copies stuff from Wikipedia -- and so the other users are no more required to attribute anyone than Wikipedia itself is.
The only way I can think of to cleanly reconcile this with the FDL is to allow authors of derived works to credit "Wikipedia" in lieu of the actual authors and copyright holders -- that is, make an explicit exception to the GFDL's requirement of acknowledging 5 authors, and allow this one pseudo-author instead. But that would require an actual license change, which at this point is probably infeasible. We could do it informally (say "crediting Wikipedia is good enough for us"), but technically any individual Wikipedia contributor could still decline to give up their right to be individually credited, since it's not actually in the license, and Wikipedia has no legal claim to the content it hosts on its website (beyond being just another licensed user). With online sources all this isn't so much an issue, since just a link back to the page resolves everything, but if anyone wants to print things in a book, the issue of whether they can credit 'from Wikipedia' or instead have to say "authored by Blah Blah, Person two, Person Three, etc." for each individual article might be a major one.
-Mark