On Tue, Sep 7, 2010 at 9:50 AM, Platonides Platonides@gmail.com wrote:
enwiki has a total of 858979 local files which sum 229 GB (and there's still commmons). 2357967 unique images (37050694 uses) are in their articles. Assuming 20Kb per image thumb (is that a good value?), that's 48 Gb, more than the 31.9 GB of the (really compressed) pages-meta-history.xml.7z but we would need to agree. They would tie at 14 Kb.
Even if all thumbs were unrealistically small, 1Kb each, they would still be several GB.
Comparing the size to pages-meta-history isn't all that fair since with the images they wont change much, so you only need to do the base copy then on the next run you just need to update/add the appropriate ones that have changed/been added or delete the ones that are gone.
Also does that figure take into the fair use images which we wouldn't be able to dump?
-Peachey