[13:06:30] ragesoss: I was gone last night, but definitely file a bug in phabricator. It could either be a database issue where the database lost / gained data since last time you ran it. It's unlikely a wikimetrics issue because we haven't deployed any real changes to the number crunching in a long time. [14:47:59] halfak: I'm making some of the case for wikibase! [14:50:21] thank milimetric. I'll file it a couple hours from now when I settle in to work. [14:50:57] cool, thanks ragesoss [16:55:35] o/ YuviPanda [16:55:46] looks like my push for pawsbase failed in the short term, huh [16:56:00] I don't see any new registrations or edits since my demo [16:57:07] hi halfak [16:57:09] hi halfak [16:57:15] hi YuviPanda [16:57:16] :) [16:57:42] hi halfak! [16:57:44] on the way to berlin? [16:57:49] yup [16:57:57] Hanging out in ORD [16:58:15] halfak: oh, is that not chicago? [16:58:27] It is Chicago. lol [16:58:37] It's my connection [16:58:41] halfak: haha [16:58:41] Are you in the airport? [16:58:52] halfak: nope, still hacking on stuff [16:58:55] kk [16:59:20] Any word from the metadata people -- did they look at my demo and just go WTF? and get back to manually editing JSON? [16:59:58] * halfak did not think he would get such blank stares from metadata people when talking about mechanisms for evolving schema. [17:00:12] * halfak does not consider himself a metadata person [17:00:14] halfak: I was at an infrastructure breakout yesterday, so not sure about that. but today they've mostly been debating what fields go where, and have apparently punted on where it'll live / how it'll be organized for later. [17:00:23] err, what fields exist [17:00:43] Stop debating! Make properties and then try to describe datasets! [17:00:59] Oh well. I lost that argument last time too. [17:01:18] IMO, designing for imagined generalities is a lot like premature optimization. [17:01:42] halfak: no arguments from me there :) I spent most of my time automating setup & teardown of jupyterhub clusters. [17:01:54] Cool :) [17:02:02] halfak: I think the largest-scale success of the 'make properties and try' is wikidata vs smw I think [17:02:23] +1 [17:03:05] Designing the whole schema first incurs a bunch of YAGNI and bike shedding. [17:03:23] halfak: I think this maybe is non-academia having tainted you :D [17:03:49] lol yeah I suppose that is probably true. [17:04:03] I don't find myself in conflict though, so I like the term "growth" [17:04:15] halfak: multiple dimensions, I think [17:04:21] but yeah, I agree [17:05:08] halfak: did you show them the reasonator UI? [17:05:17] I find that usually helps me whenever I show people wikibase [17:05:29] I didn't. I wasn't exactly showing them wikidata. [17:05:49] But I did show how we can tie into wikidata and I talked about efforts to describe research manuscripts in wikidata. [17:06:20] If wikidata has authors & manuscripts, then pawsbase can bring datasets and code. [17:06:33] code --> source? [17:06:36] Not sure about that one [17:06:43] halfak: right, but I think reasonator + wdqs equivalents are super important bits of it, for querying + exploration. [17:06:56] from vague conversations I think they didn't quite realize you could do all those? [17:07:12] I pointed at wdqs and said it does SparQL. Then I said, I don't grok SparQL so no demos [17:07:15] :S [17:07:18] ah [17:07:37] halfak: but you're now listening to third hand info from me, so might not be entirely accurate. [17:07:54] halfak: since it's me talking about what I inferred people to be talking about yesterday from listening to them talk today [17:14:54] I think I might need to take sit down with one of the metadata pros and give a more clear argument for wikibas.e [17:15:14] Hopefully we can get an advocate that speaks the language & has the cred. [17:16:16] halfak: yeah [18:36:55] halfak: I wonder if this is a generally useful application: 1. given a JSON Schema, 2. generate a UI for it that 3. allows people to use it with wikibase [18:43:37] halfak: well I tried to help you on https://phabricator.wikimedia.org/T134651#2323991 but I fear it will take some shoulder tapping [18:43:44] to get the right folks [18:45:41] chasemp: I think it is wes for halfak [18:46:07] any c-level will do ironically, I think luis did it back when as he was "around" [18:46:24] and erik [18:46:39] ya [18:47:08] actually [18:47:10] https://phabricator.wikimedia.org/T133145#2318733 [18:47:15] https://phabricator.wikimedia.org/p/Wwes/ [18:47:18] he has done it :) [18:47:31] yup! [18:48:53] I'm not really following that request to know if it's ready but if is I would imagine halfak ping @wwes [18:49:22] chasemp: yup, it is I think. halfak is in a flight atm, will probably check in on after