Re: [xwiki-devs] Improvements to L10N to see a quick status by language

Hi, I can try to look into this. What we can do I think is: 1/ Move to custom mapping of the translations class - requires migrating the data - requires rewriting the queries use throughout the application This should speed up all calculations. 2/ Move counting of missing translations in it's own data set and use a scheduler job to recalculate regularly these counts - requires informing the user that data might not be fully up to date, and eventually put a button to recalculate for a specific language/project - the best contributor ranking would not necessarly need a refresh and would only need to be calculated from time to time - we can use this opportunity to display some full counts of missing translations for a country/project including a rnaking by country which is not doable right now I believe that these are the most costly operations at this point. Moving towards pre-calculated data would speed up things incredibly. If that seems ok for all I would do the following: - take the L10N app locally and setup custom mapping - rewrite the queries for both real time calculation and scheduler based calculation - migrate L10N by taking it offline half a day - create the new classes to hold the pre-calculated data - implement the scheduler job to preform the calculation and save to the new classes - implement the queries to display calculations from the pre-calculated data - add some new queries to get a full language ranking WDYT ? Ludovic 2013/1/18 Vincent Massol <vincent(a)massol.net> -- Ludovic Dubost Founder and CEO Blog: http://blog.ludovic.org/ XWiki: http://www.xwiki.com Skype: ldubost GTalk: ldubost