On Mon, Feb 23, 2009 at 2:44 PM, David Gerard <dgerard(a)gmail.com> wrote:
2009/2/23 Newyorkbrad (Wikipedia)
<newyorkbrad(a)gmail.com>om>:
However, one question that I have is whether the
dump includes, or should
conclude, all namespaces, or only articles. In the past, there have
allegedly been instances in which database dumps have been utilized for
purposes such as harvesting oversighted edits in userspace and utilizing
the
information for purposes of harassment. I am not
sure whether there is
value to providing dumps of other than the content spaces. Comments?
The value of providing good dumps is forkability, in case WMF is hit
by a meteor, hit by a legal meteor, goes collectively insane, etc.
Imagine trying to fork Wikipedia without being able to take the
project spaces with you.
You mean we haven't already gone collectively insane?
To answer Brad's original question: different dumps contain different
information. There's the article-only dump that most mirrors, etc. use
and there's the larger full-wiki dump. It's the latter that is most prone
to failure and tends to kill the overall dump process.
-Chad