[02:08:46] Tim-away: Wouldn't it be easier to use .splt() on rawDictionary and then use arrayDictionary.indexOf( p[0] ) instead of looping through and going into an object and checking p[0] in hashtable ? At first it made sence since "key in object" is broader compatible than indexOf() for old browsers, but since indexOf() is used anyway.. [02:11:22] arrayDictionary would be rawDictionary.split( ' ' ) [02:12:10] that would be slow [02:12:30] hashtable lookup is O(1) [02:24:16] Tim-away: ah, indeed. Much slower. [02:25:06] maybe searching rawDictionary directly with indexOf would be better [02:25:20] makes sense, lookup can be done right away, indexOf needs more searching. Especially when done in the user interface directly one can notice the difference [02:25:42] then you'd avoid the setup overhead, which is probably the bulk of the average running time since people don't type many characters into that box [02:25:44] but adding a space before and after the rawDict and checking index of (space)word(space) would be quicker [02:25:54] yeah, that ^ [02:26:36] though perhaps that's not even needed, depending on how the dictionary. [02:28:48] according to my profiler (WebKit, Safari 5 : Mac, Core2duo laptop) checking rawDictionary.indexOf(val) for each of the 8509 words takes 0.9 seconds. [02:28:57] checking val in hashtable takes 0.88 [02:29:40] since the difference in factor 8509 is 0.02 seconds, I think in actual use the difference is unnoticable [02:29:54] but whatever, it's a pretty handy script :) [02:31:09] indexof on a string is almost as fast as object lookup. indexof on an array is slow no matter what. [02:37:50] oops, that's the british english dictionary [02:38:14] there should be 8531 [02:38:20] I'll have to update it [02:41:52] ;) [02:55:33] updated [02:56:43] Anyone know where the entry "List of 2006" is coming from here ? http://en.wikipedia.org/w/api.php?action=query&list=allcategories&acprefix=List%20of [02:56:55] there doens't seem to be an article or category by that name, not deleted either. [02:57:14] http://en.wikipedia.org/wiki/Category:List_of_2006 [02:57:31] (same goes for a few others on that list) [04:07:30] mysql> select * from category where cat_title = 'List_of_2006'; [04:07:30] +--------+--------------+-----------+-------------+-----------+------------+ [04:07:34] | cat_id | cat_title | cat_pages | cat_subcats | cat_files | cat_hidden | [04:07:37] +--------+--------------+-----------+-------------+-----------+------------+ [04:07:40] | 599828 | List_of_2006 | 0 | 0 | 0 | 0 | [04:07:44] +--------+--------------+-----------+-------------+-----------+------------+ [04:07:47] 1 row in set (0.06 sec) [04:07:53] Looks like an orphaned row in the category table. [08:45:00] I would like to add 2 keywords to bugzilla: captcha and analytics, can somebody do this or give me the user permissions so I can do it myself? [21:46:04] [03:49:44] I would like to add 2 keywords to bugzilla: captcha and analytics, can somebody do this or give me the user permissions so I can do it myself? [21:46:11] I think CAPTCHA bugs are going to use a tracking bug. [21:46:19] Though if you really want new keywords, talk to ^demon. [21:46:39] Analytics would probably be a component or something. [21:46:48] <^demon> Sounds like a component. [21:47:40] keywords are fairly high-level. I'm not sure how many competing analytics solutions we have. It's gotta be limited to one or two extensions, I'd imagine. [21:47:50] analyics would be a keyword for bugs that can exist in xml snapshot / wikimedia downloads / and mwdumper export /import [21:47:50] WMF is playing with OWA, but I'm not sure what the status of that is. [21:48:08] owa would also fall under analytics [21:48:10] Hmm, then "analytics" is probably the wrong word tehre. [21:48:12] there [21:48:16] offline? [21:48:31] so these are conceptually related feature requests / bugs [21:49:20] <^demon> tracking bugs are nicer than keywords, imho. [21:49:54] but tracking bugs are (AFAIK) used to track the progress of closely related / interdependent bugs [21:50:11] that is definitely not the case for analytics, maybe it is for captcha [21:50:52] with a tracking bug you don't get an overview in a single page of the status of each bug, while with keyword search you do