Jimmy Wales wrote:
First, blogs-as-sources is already a tricky topic, but there are of course cases where a blog is a legitimate source. For example, if a well known person blogs in response to a media controversy, that particular blog post can be quite valid as a source for a sentence saying "In a post to his personal blog, John Doe vigorously disputed the allegations put forward by the New York Times."
Now suppose that same well-known person, in a completely different matter, gets into some kind of squabble with a Wikipedian and uses their blog as a weapon in that dispute. In some extreme cases (death threats? libel? we could discuss...), there could be a reason to delink the blog everywhere.
I don't see this being a reason even in extreme cases. If the person in question posts a death threat or libel in some other completely unrelated blog posting, how does that in any way affect the fact that he "vigorously disputed the allegations put forward by the New York Times" in the original blog posting being used as a reference?
Or, in case of a redirect to an attack page, there is absolutely a reason to delink the blog (because the link is no longer valid).
That's already covered by existing guidelines. The procedure when a reference URL changes to something other than the referenced text is to try to find the old version in archive.org, or otherwise "fix" it. If it can't be found anywhere, though, we still don't remove the reference but rather record the date that the original link was found to be inactive; even inactive, it still records the sources that were used, and it is possible hard copies of such references may exist, or alternatively that the page will turn up in the near future in the Internet Archive, which deliberately lags by six months or more.
http://en.wikipedia.org/wiki/WP:CITE#What_to_do_when_a_reference_link_.22goes_dead.22
I suppose if the "dead" link goes to something actively misleading instead of just a standard 404 page one should deactivate the link, though, to make sure readers notice the fact that it's not the material that was originally referenced.