[01:38:45] * halfak is finally writing some code. [01:38:48] It's been ages! [01:39:19] (running R snippets to check the math in papers doesn't count as "writing code") [01:39:29] IMO [02:42:14] I have streamed some XML. It was good. [16:59:00] Hey Nettrom, I added you as collab @ https://meta.wikimedia.org/wiki/Research:Measuring_article_importance [16:59:54] How do you want to move forward? [17:02:03] * halfak runs off to grab food. [17:02:04] brb [17:15:00] * halfak is back [17:15:09] Oooh that rhymes :) [17:15:26] halfak, bustin' da rhymes [17:15:30] better watch your back. 'cause here comes the halfak [17:15:49] I see you and Ironholds showing off your skills at future WMF parties [17:15:59] :D [17:17:07] Nettrom, re. importance measures, how do you want to move forward? [17:17:22] halfak: good question, I went through my results from earlier this year [17:17:30] arguably, they need to account for redirects to be better [17:17:58] I'd also like to see if there's additional literature that argues either way for importance [17:18:25] e.g. there was a paper at HyperText this year that argues similarly as the Group '07 paper [17:19:55] what does your timeline look like, how quickly should the page be fleshed out? [17:20:32] Personally, I just want to get a dumb system up and running ASAP, but I think that this bigger question about importance is interesting. [17:20:46] I'd like to add interesting tid-bits to the article as I come across them. [17:21:03] Like that HyperText paper. [17:21:24] And the regular analysis I'll need to do to get a dumb version of WikiCredit working. [17:21:42] dumb == MVP [17:22:11] In this case, MVP == I have not worked out all of the implications of the measurements I am taking, but they are more good than bad. [17:22:24] mhm [17:24:11] There's a guy I worked with at Wikimania who was particularly interested in this question of Importance. [17:24:22] halfak: by dumb system, are you thinking some type of service that'll say something about the importance level of an article? [17:25:05] Rather, that WikiCredit will just use some measure (e.g. log(page_view_rate) or log(inlinks)) [17:25:21] mhm [17:25:32] btw, I have a plot somewhere of views vs inlinks, I think [17:26:39] oh, that's only for the vital articles [17:27:23] "vital"? [17:27:34] yes, there's a WikiProject about vital articles, I think [17:27:38] Gotcha [17:28:03] https://en.wikipedia.org/wiki/Wikipedia:WikiProject_Vital_Articles [17:29:32] That seems like it would be interesting. [17:29:36] hm, some interesting stuff here: https://en.wikipedia.org/wiki/Wikipedia:Village_pump_%28miscellaneous%29/Archive_37#Can.27t_decide_what_to_work_on.3F_Try_a_short_popular_vital_article.21 [17:29:39] For the Article importance page [17:30:12] BTW, you've got access to my google docs about pageviews and importance, where importance is flagged by the WikiProjects [17:30:49] (although not with your WMF account) [17:31:18] No worries. Old one still works :) [17:40:25] * Nettrom food [18:58:31] * halfak gets on bike to join other MN WMF folk at a coffee shop [18:58:35] (actually a tea shop) [21:52:03] Ironholds & leila: I've got a process to run on stat3. Looks like you guys are done. Is that right? [21:52:22] halfak, Ironholds is done, I'm not, but you can go ahead and run yours [21:52:29] Great. [21:52:31] er [21:52:33] I'll need another hour to prepare few things. [21:52:34] not quite done yet! [21:52:45] but my memory usage is much lower than it was [21:52:45] I don't need any meaningful amount of RAM [21:52:52] And like 25% of a core. [21:53:00] cool! that's totally fine, halfak [21:53:07] thanks for checking [21:53:27] :)