[Foundation-l] Long-term archiving of Wikimedia content

[Anthony]

On Sun, May 10, 2009 at 5:25 PM, Casey Brown <cbrown1023.ml at gmail.com>wrote:

> On Sun, May 10, 2009 at 5:16 PM, Anthony <wikimail at inbox.org> wrote:
> > Mostly I meant the user data (especially the passwords).  The relative
> value
> > of them compared to the rest can be shown by anyone who tries to create a
> > fork.
>
> In the dumps, these are always done first:
> <http://download.wikimedia.org/enwiki/20090506/>


Pretty much in order of importance (except that I don't know what "Update
dataset for OAI updater system" means).

I suppose the stub dump is also important as it's not widely replicated and
contains the author information, but in a month or so the GFDL is going to
be dropped anyway so that wouldn't be such a huge loss I guess.

In any case, the Library of Alexandria analogy kind of forgets about the "no
original research" policy.  Even if the current versions of all articles
were lost, it wouldn't be all that horrible.  It might even be a good
thing.  Consider Citizendium's decision not to go the route of the fork but
instead to start from scratch.