User talk:MikeMol: Difference between revisions
→Rendering speed
(→Is file uploading blocked forever?: Uploads.) |
|||
Line 565:
(I suppose [http://blog.wikimedia.org/2014/12/29/how-we-made-editing-wikipedia-twice-as-fast/ wikipedia's approach] might work, but I don't know what all efforts that migration involves...) --[[User:Rdm|Rdm]] ([[User talk:Rdm|talk]]) 04:38, 10 May 2016 (UTC)
: It's getting to be a real problem. The recent issues with session lost were driven by storing page rendering cache data in the same place as the session cache. Session cache is now in the DB, but, yeah, definitely need to start moving on this. I really would love to be able to migrate to a semantic-assembly model like we attempted a few years ago, we ran into bugs / limitations in Semantic MediaWiki when we tried it. Longer term, I'm seriously thinking MediaWiki is aging out as the servicable tool for the job. It imposes constraints that are quite inconvenient (seriously, we should be able to say "show me all the examples written in C" or "show me all examples of this task written in a functional language"), and then there are page rendering issues like these. --[[User:Short Circuit|Michael Mol]] ([[User talk:Short Circuit|talk]]) 17:02, 8 August 2016 (UTC)
:: Any implementation is going to have issues. Conceptually, we should be able to modify the mediawiki code (first as a specific hack for our needs, then redoing it in plugin format and engaging whatever is left of the mediawiki team - or maybe just forking the code). I think it's mostly php, which is ... ad hoc, and more than a little unpredictable... But unless we are willing to write a replacement ourselves, I don't think we're going to find anything which is particularly suited to our needs. That said, I'd love to be proven wrong on this. --[[User:Rdm|Rdm]] ([[User talk:Rdm|talk]]) 17:09, 8 August 2016 (UTC)
| |||