An'n 11.10.2010 20:13, hett Strainu schreven:
2010/10/11 Marcus Buck<wiki(a)marcusbuck.org>rg>:
There was a Google Summer of Code project:
<http://www.mediawiki.org/wiki/User:Peter17/Reasonably_efficient_interwiki_transclusion>.
It's basically ready to use. About the _actual_ implementation you have
to ask the Foundation developers.
Marcus, I would hardly call that project
"ready to use". It leaves
many issues unresolved, such as:
1. local editing of the remote data with unified/non-unified accounts
2. automatic translation importing from translatewiki (people would
probably want to use localized parameters/template names)
3. all the known limitations noted there :)
It looks like a good start, but I somewhat doubt we will be seeing it
in production soon.
If in Nikola's solution all this works, I wasn't aware
of it. #1 to me
actually seems like an advantage. If data is changed for all wikis users
must go to the central wiki to edit it. Otherwise it'll definitely lead
to problems. #2 also is only a problem if we accept that #1 is wanted as
a behaviour.
Actually I have no specific preference for any of the two solutions. I
just wanted to hint at an alternative effort.
The only thing I care about is, that _some_ solution is found and
implemented. Both solutions can be implemented in a short period of time
if only somebody cared to start the process. It's the most important
development step for Wikimedia in years. Possibly ever.
See e.g.
<http://lists.wikimedia.org/pipermail/foundation-l/2010-August/060628.html>.
Marcus Buck
User:Slomox