Alex756,
The history is not a great concern. It's
Since we can't know for sure that rationale
is:" probably" better to delete "potential" copyright infringements <
If the article had gone through many revisions and much of it was not infringing it would be less clear that deletion was a good course. Losing the history with a delete and creating the article from the pieces which didn't infringe would lose the history of the non-infringing gradual construction of the page and risk losing that defence against a bogus infringement claim based on accidental resemblence to some other document. Not such an issue in this case because the history is insignificant. I don't want to see delete as the method for all cases, though.
To see how Google handles this, take a look at http://www.google.com/search?num=50&hl=en&lr=&ie=UTF-8&oe=UT... and follow the convenient link they provide to the takedown notice, which includes the content they have been asked to take down, effectively making it available again. That's approximately what the history pages here do, though I agree with your archive argument and suggest also that the history is a working document and not part of the current publication, which is the Wikipedia itself, not the history of every page in every version. That's well illustrated by the many sites making the Wikipedia but not the history available.
Sounds as though you and I and Google agree that it's not likely to be problematic (you wrote "probably no damages or no infringement or both") and the only potential disagreement is in how far to go to reduce the chance of a suffering a lawsuit which will fail on its merits but still cost money. Personally, I'm inclined to think that anyone taking action against the Wikipedia will generate strongly negative publicity for themselves and that the risk of that will significantly inhibit baseless legal action. Besides, it might be cheap publicity as long as its settled in the very early stages ... AFTER the Wikipedia press release and Jimbo Wales TV appearances about it.:) Settlement offer along the lines of "$1 as far more than your demonstrable losses and you don't have to explain to the court and the ethics committee why you're wasting the resources of the court".
Still, if Jimbo is more comfortable with a delete, he's the one other than the contributor who is most likely to be hit by legal action and his comfort level matters so long as that remains true.
It might be useful to downstream users to keep a clear log of what we've acted on to deal with actual copyright infringement notices, to help them know tha thtey have pages which are likely to attract legal attention. We're also lacking, so far as I know, a page listing the people we've contacted for permission and what the result was, so we can avoid making repeated requests for the same thing. Think its worth creating them?
Yes, I've read the history page discussion. I generally agree with the view expressed by Brian Vibber, that anyone may choose to use any revision of the Wikipedia they wish, or any combination of revisions and derivative works based on them. That's not inconsistent with the intent to have the current Wikipedia as the currently published document. The history is really just our working document available as part of the process of constructing the actual work.
Where I disagree is in cases like abuse or personal information about themselves which private people contributing to the Wikipedia (rather than public record about public figures) want removed. While it's true that they were once published that way, it's socially good to assist people with removing such things, including ceasing to make them available through the history. Yes, someone could have the page and it would be GFDL, but it's still socially good for Wikipedia to stop distributing it, to inhibit further distribution. Our removal also makes it a bit easier for them to pursue a right of privacy action against whoever is harassing them, though I wouldn't like to speculate on their chance of success.
I don't agree that talk pages or the mailing lists need to be under the GFDL. I think that gets in the way of frank discussion and isn't necessary. A narrower licence, "GFDL but only within the Wikipedia project and for the purpose of that project", suffices for them and doesn't get in the way of researchers, who have ample fair use rights for the purpose.