On 1/29/07, Stan Shebs stanshebs@earthlink.net wrote:
Hee hee, I'm thinking I should take a couple weeks of wikicommonsbreak right about now! But seriously, what you're talking about will stir up a large percentage of the commons people who obsess over categories. I'm not even sure having category intersections in the software will satisfy some of these folks - they go on about server loads and the like as well. I'd suggest just bringing up the issue and watching the fur fly, before deciding to invest time in rewriting anything...
:) But the fun is just beginning! :)
As far as load goes, it's only a question of implementation. A few months back I posted some example performance data for intersections using inverted indexing using the actual enwiki category data: http://lists.wikimedia.org/mailman/htdig/wikitech-l/2006-September/026715.ht...
Even the most evil cases (intersecting two huge categories) ran very quickly. (For example the intersection of GFDL images and living_persons each which has over 100,000 members took 25ms).
The technical obstacle of getting the software implemented is far smaller than the data quality issues we have.