[09:27:38] Curious jump compared to other months http://stats.grok.se/en/201509/Hugh_of_Saint-Cher [15:42:57] Hello, scientists. [15:53:22] * Emufarmers feels ungreeted. [15:57:36] * guillom hugs Emufarmers. [15:58:42] :> [16:00:57] o/ Emufarmers et al. [16:00:58] :D [16:07:43] Time zone-appropriate greetings, scientists and hangers-on. [16:08:48] It's distribution-upgrade day! [16:13:53] guillom, which distro? [16:14:01] * halfak is still on Ubuntu 14.04 [16:14:23] I might be due for a dist upgrade, but I don't want to deal with the potential problems :\ [16:14:32] I've already solved all of the critical problems with this version! [16:14:36] * halfak feels like a sysop [16:15:26] halfak: https://news.opensuse.org/2015/11/04/opensuse-leap-42-1-becomes-first-hybrid-distribution/ [16:15:41] With KDE Plasma 5 \o/ [16:16:26] How do you install packages with that distro? [16:16:47] zypper (CLI) or YaST (GUI) [16:17:57] Interesting. I'm not sure I've ever seen a Zypp package [16:43:40] halfak: hey, assign something to me [16:43:41] :) [16:43:52] oh let's go to #wikimedia-ai [16:44:24] Morning halfak! I have a question :D, do you know any wiki project IRCs that are particularly active right now? [16:45:50] Hey BeiYan. Sorry, i don't have any good recommendations. [16:46:11] Oh wait [16:46:21] #wikiprojectx sees a lot of activity [16:46:37] halfak, thank you so much! [16:46:58] * halfak takes Amir1 up in #wikimedia-ai [18:30:50] I botched my upgrade /o\ [18:37:38] guillom, nooooo [18:37:44] Hate when that happens. [18:40:56] "No space left in device." I used the option to download all the packages in advance, to avoid network issues, but now the installation process is missing space to do the actual upgrade. I'm going to need to do a fresh reinstall. Not necessarily a bad thing; I should go home anyway because allergies. [18:41:20] * guillom will do that after metrics meeting and food. [18:47:04] Always a good idea to stick around for the food. [18:47:58] 13 minutes, yes? [18:48:26] indeed. [20:03:45] guillom: on what distro? [20:04:26] With fedup, for instance, AFAIK nothing actually changes before the reboot, so an interrupted download shouldn't cause anything bad. [20:06:59] Nemo_bis: opensuse. The machine doesn't even go past boot because there's no free space for startup. [20:08:06] ugh [20:08:21] guillom: how about a live boot and free up some space? [20:29:10] Getting on bike. Back in time for Research Fellow meeting of doom [20:29:12] o/ [21:02:56] Nemo_bis: yes, I'll try a chroot before reinstalling. [21:15:25] o/ [21:15:29] * halfak is back [22:34:46] halfak: I just got your email! [22:34:48] * YuviPanda reads [22:35:08] I'm not quite sure what they are looking for, but I'm sure we can use our imagination. [22:35:24] halfak: woah, the proposal was accepted! yay! [22:35:34] What!? Did I not CC you on email [22:35:37] halfak: unfortunately I also found out that the workshop is in like the last days of Alice bineg in SF [22:35:40] which is sad :'( [22:35:43] but I'm sure we'll manage [22:35:45] halfak: no! [22:35:49] Workshop is in SF [22:36:12] :( I'm sorry for the oversight. [22:37:42] np :) [22:37:52] I should've checked before she booked her tickets... [22:38:02] halfak: but this means I might not be around during the event itself... [22:38:08] although I can pop by for a couple hours now and then [22:38:10] err, the main event [22:38:14] I'll be there during the workshop [22:38:21] just for... other stuff? [22:38:22] Yeah. I think that is OK. [22:38:36] yeah, we can flag things I'll be super interested in and negotiate some time around that :) [22:38:44] I'll also see if I can move her flights around :) [22:38:49] We will probably need you to be around to (at least) sysop the workshop [22:39:01] I can help you find interesting things that'll be worth your time to attend at CSCW :) [22:39:09] I'll be around fully engaged for all the workshop, sure. [22:39:18] +1 [22:39:34] So anyway... about this metadata interface. [22:39:46] right so I don't fully understand what it is and what it means [22:39:49] These guys have a semantic mediawiki that they have been using for metadata. [22:39:54] oh dear god. [22:39:56] !hss etc [22:39:58] Yeah. [22:40:05] "!hss"? [22:40:13] in #mediawiki in olden times [22:40:16] !hss replied with' [22:40:20] HE SAID SEMANTIC, GET HIM! [22:40:24] Oh! [22:40:35] in reflection of the cordial relationship between the SMW and MW devs :) [22:40:53] annnyway [22:40:56] Yeah. So anyway, not wonderful, but they are working with it and we can probably extract the data later if they want to migrate [22:40:57] what does this metadata look like? [22:41:02] right [22:41:06] is there a link to this wiki? [22:41:18] I don't see it in the link [22:41:18] * halfak tries [22:41:20] err [22:41:22] in the email [22:41:52] http://wiki.urbanhogfarm.com/index.php/Main_Page [22:42:09] http://wiki.urbanhogfarm.com/index.php/Google%2B_Circles [22:42:43] Looks like this one is in their new structure: http://wiki.urbanhogfarm.com/index.php/3D_Manufacturing_Community_Surveys [22:42:58] Looks like it all key-value pairs. [22:43:15] http://wiki.urbanhogfarm.com/index.php/3D_Manufacturing_Community_Surveys [22:43:17] haha [22:43:19] I'm there [22:43:21] now [22:43:26] lol at course name [22:43:35] what's the earliest date thing? [22:43:39] Documentation: "Yes" redlinked [22:43:51] Must be the earliest date in the temporal data [22:44:02] ah so keys depend on the type chosen [22:44:25] I think so. [22:44:43] is there docs on what 'temporal' means [22:44:51] Presumably [22:45:06] http://wiki.urbanhogfarm.com/index.php?search=Temporal&title=Special%3ASearch&go=Go [22:45:07] No [22:45:09] lol [22:45:30] It's quite possible that this co-design process could kick Kristen into gear [22:45:37] Because this wiki needs some content [22:45:53] Apparently they have classrooms full of students working on this. [22:46:07] what is 'this'? the wiki? [22:46:11] Yeah [22:46:19] Specifically describing datasets on it [22:46:26] Anyway, I hope they'll clarify what type of integration they have in mind. [22:46:46] so what I'd like is a YAML file or something that Quarry can directly interpret to load into memory [22:46:51] well, a script can interpret to load into memory [22:46:52] And independently or in parallel, we can think about the kind of metadata integration we think quarry should ahve. [22:46:54] err [22:46:58] not in memory, but into mysql [22:47:02] and then it'll have a link to the YAML file [22:47:05] which we can display somehow [22:47:41] * YuviPanda isn't tied to YAML, just a hierarchical tree-like thing [22:47:43] I wonder if we can leverage wikibase or smw to provide the editing interface and use the API to request data. [22:48:20] smw because they already have it or wikibase because it's better [22:48:27] on a personal level I'd like to stay away from SMW if possible since it's given me nothing but pain... [22:48:36] Yeah. I agree. [22:48:41] not sure about wikibase but I do know it has like 400 million dependencies... [22:49:07] Hmmm... Harej's got one running. [22:49:28] on labs-vagrant I think? [22:49:28] Not to say harej isn't a smart dude, but maintaining live systems isn't his day job. [22:49:29] Is it? [22:49:34] We could do that :) [22:49:45] My body is a live system. It's dubious how well I maintain it [22:49:49] Should be doing more experimentation than optimization now anyway [22:49:52] Composer handles all the dependencies [22:50:20] harej, live, but is it also adaptive? [22:50:23] You download Wikibase from github.com. You use composer to download the rest. [22:50:33] I don't know how adaptive my body is. [22:50:55] Either way, I think we should decide how we want users to interact with metadata before we settle on a technology. [22:50:58] J-Mo, ^ [22:51:08] Talking about setting up for workshop and metadata in quarry long term [22:51:53] nice! [22:51:57] If I could have exactly what I wanted, it would be a tabular data wiki with quarry as a front-end. [22:52:24] Kristen just sent out an email about getting metadata set up. [22:52:31] I'll CC you quick J-Mo. [22:54:03] cool [22:54:44] hmm [22:54:57] so I vaguely feel like I don't really fulllly understand what's needed [22:55:22] using a separate WIkibase (or anything MW really) always feels like 'mowing lawn with tank' to me, but I'm biased [22:55:41] I'll admit to being still slightly in teh dark about how this would be used [22:55:50] is it just a place for people to browse and edit? [22:56:09] and then jump off to do other things with? [22:56:40] YuviPanda, good questions that we should make Kristen answer. But I do have my own thoughts ... [22:57:08] right, so I do feel like there's a lot of 'implied context' that you guys have that I'm missing :D [22:57:16] so what I'd like is [22:57:26] 'in a world where this wiki/whatever does not exist, what would change?' [22:57:26] Even for the tables that are a standard part of MediaWiki, I want a collaboratively edited space where we can capture description, historical anomalies, discussion, etc. [22:57:30] how would you do the things you would want? [22:57:53] (if this didn't exist) [22:58:01] oh yeah, totally [22:58:13] It would be great to access that from some sort of catalog in quarry (could be just a predictive search box that opens collaborative docu-thingie in a new tab) [22:58:42] I'd want to be able to search for things like "Datasets with users and timestamps" [22:59:15] Or "why is user_registration NULL for all the users who registered before 2006 in every Wiki except enwiki?" [22:59:29] ah I see. [22:59:49] so you basically want structured info about particular datasets, including but not limited to schema [22:59:56] Maybe even "Datasets that have something to do with 'quality' or 'productivity'" [23:00:02] Yes [23:00:30] A structured wiki with a talk page does seem like the right abstraction. [23:00:52] Yeah. I can see how I would use it. But the tank lawnmower feedback is well received. [23:00:57] Simple is awesome if we can have it. [23:01:10] Not needing to think through permissions and contribution patterns is nice too. [23:15:32] halfak: yeah, totally. [23:15:49] halfak: it also depends on how much cross linking we want vs just 'free search' [23:15:57] if the latter we can get away with a simple contenthandler [23:16:09] (like the Schema: pages on meta) [23:16:37] Yeah. I think that might be a good thing to test at the workshop [23:16:50] right, so how do we test it? [23:16:50] We could go with an easy option and see if people *want* something more complicated. [23:17:20] right. so 'easy option' would be, IMO: just a YAML editing interface (do we want schema validation for that?) + a flow talk page [23:17:45] Think we can dump that on a live wiki right now? [23:20:32] as in? [23:20:35] a pre-existing wiki? [23:20:52] I can probably spin up a vagrant on labs setup pretty quickly [23:33:04] harej: how was your experience running wikibase? [23:33:16] set it up once, didn't really have to touch it since [23:33:28] YuviPanda, was thinking of using Meta or MediaWiki to host the flow pages or something like that [23:33:36] though! pay attention to the file that specifies the blocked out item numbers [23:33:36] Sorry. 60% in a meeting [23:33:37] ;) [23:33:51] harej: no with actually using it? [23:33:55] No Q666? [23:34:14] Well, you can edit which item numbers are blocked out, for later database manipulations or whatever. But it prevented me from creating a Q1