Re: [Wikitech-l] image bundles' size?

2010-05-17 Thread Daniel Kinzler
James Salsman schrieb:
 How large would the projects' image bundles be uncompressed, if they
 were to exist?
 
 Also asked at:
 
 http://meta.wikimedia.org/wiki/Talk:Data_dumps#How_big_would_image_bundles_be_if_they_existed
 
 However, someone suggested I should be on wikitech-l more, so I
 thought I would try asking here.  I read here regularly, but I prefer
 the dogfood.  I promise to send the answer to the other place the
 question was asked if someone else doesn't do so first.

AS I said there:

Enwiki: 200 GB, Commons: 6 TB. roughly.

-- daniel



___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l


[Wikitech-l] image bundles' size?

2010-05-16 Thread James Salsman
How large would the projects' image bundles be uncompressed, if they
were to exist?

Also asked at:

http://meta.wikimedia.org/wiki/Talk:Data_dumps#How_big_would_image_bundles_be_if_they_existed

However, someone suggested I should be on wikitech-l more, so I
thought I would try asking here.  I read here regularly, but I prefer
the dogfood.  I promise to send the answer to the other place the
question was asked if someone else doesn't do so first.

This is for a mirror project, which I want to fork into a peer-to-peer
wiki system.  A serious problem with peer-to-peer wikis is edit
conflict resolution -- most everything else about syncing is not as
hard as that in general.  People often make the mistake of using git
as a metaphor, but code merges are much more tightly coupled with the
text being edited than most Foundation projects.  But edit conflict
resolution is so hard in peer-to-peer mode; for example, the two
editors in conflict may be unavailable, and the person faced with the
conflict may not understand the original or either of the two edited
versions. We can use croudsourcing systems to, for example, have three
people try to resolve each nontrivial conflict, and three more people
to decide which of the first three was best; if there isn't
substantial agreement, get a fourth proposal in light of the first
three, etc.  Can anyone think of a way to motivate volunteers to
resolve edit conflicts as a third party?

Sincerely,
James Salsman

___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l