WOS4: Quality management in free content [LWN.net]

WOS4: Quality management in free content

Posted Sep 20, 2006 17:52 UTC (Wed) by zooko (guest, #2589)
Parent article: WOS4: Quality management in free content

Sounds like a job for a revision control tool. The Citizendium can put their copy of wikipedia under control of such a tool, then use that tool to merge their edits with wikipedia edits, as well as to include/exclude wikipedia edits on a "cherry picking" basis.

darcs would be perfect the job, were it not for its fatal performance flaws (i.e. it always locks up when you try certain kinds of merges).

to post comments

Free content and version control

Posted Sep 23, 2006 10:05 UTC (Sat) by ddaa (guest, #5338) [Link] (1 responses)

> Sounds like a job for a revision control tool.

I had just the same impression. The "hard-fork on first edit" approach just sounds like "we do not have the right tool to track upstream, so we fork".

> darcs would be perfect the job, were it not for its fatal performance flaws

Hu... and CVS would be perfect for the job were it not for its being rubbish?

I take you actually meant: there is not at this time a practical free software tool to deal with cherrypicking inclusion and exclusion, although the free distributed version control folks have been thinking a lot about it, it does not seem that anybody was able to come up with something usable so far.

Free content and version control

Posted Sep 27, 2006 5:29 UTC (Wed) by njs (subscriber, #40338) [Link]

Cherrypicking per se -- i.e., taking some changes from another version and applying them to your version -- is pretty easy. (I mean, these are just text files, and diff/patch can handle those well enough for most purposes, though you can also do somewhat better with real merge tech.) Where darcs runs into trouble, and where there are unsolved problems, is when you want later merges to be smart about respecting those past cherrypicks. But since they don't want to re-merge the two forks anyway, that doesn't matter. I don't see why you couldn't perfectly well import future edits from wikipedia into a "possible changes to review and click on the ones that you want to apply" interface.

(Some of us have made a similar argument about cherrypicking for software too, since the most important use cases are things like cherrypicking to stable branches that are never going to be merged back either.)