[00:03:36] hi [00:05:09] never mind I found the answer (it was on instance of vs subclass of) [00:11:33] I'll still verify with you [00:11:41] https://www.wikidata.org/wiki/Q82516 <- Subclass of SOC [00:11:59] Is a family of SOCs a subclass of SOC? [00:12:01] I'm not sure [00:12:07] it's on other SOCs [00:12:21] but it's not a SOC, it's a family of SOC [00:12:30] then there are more narrow family like I.MX25 [00:12:51] subclass of is the right thing, then yes [00:12:56] and then the actual SOCs (Like the industrial version with lower CPU freq for instance) [00:12:59] ok [00:13:03] what would I.MX251 be? [00:13:04] and the individual SOCs are then instance of that family item [00:13:13] ok [00:13:24] so I did it right [00:13:25] thanks [00:14:30] Then the idea is to add the hardware blocks from here, but without the freedom concerns yet: https://libreplanet.org/wiki/User:GNUtoo2/Hardware_freedom_documentation/SOCs#I.MX_25_.28Consumer_and_Industrial.29 [00:15:09] If I can add a link to the datasheet and point to its respective chapters it would be nice [00:16:24] As for the actual blocks, I'll refer to the datasheet names and cross check with the drivers to see if they are (mostly) the same block when other SOCs will use them [00:16:25] that's very well possible [00:19:57] 1) I'll create let's say an "I.MX UART" property, but then how do I link it to the I.MX25{1,3,5,7,8} ? [00:20:24] sorry [00:20:28] s/property/item/ [00:20:53] That item is documented in a datasheet (publically downloadable) [00:21:20] in that case you should link it as reference (using reference URL) [00:21:28] and it's present on the I.MX25{1,3,5,7,8} (And other I.MX too, like I.MX35, 51, etc...) [00:21:32] ok [00:21:40] but can I add the chapter reference? [00:22:02] Sure https://www.wikidata.org/wiki/Property:P792 [00:22:09] ok, thanks [00:22:27] So you add a reference url, a chapter, access date would also be nice and maybe language of work [00:22:39] ok [00:34:18] hoo: how can I say that the UART block is inside the I.MX251? [00:34:36] has part and part of [00:35:55] thanks [00:43:13] I'll add the link later [00:43:21] then I'll have to switch to the command line [00:43:30] (Editing trough a vm is really really slow) [00:43:43] bye, thanks a lot [08:55:08] Morning! [08:55:33] morning! [08:57:22] aaaaaaaa [08:57:46] Will there be a mode where all possible languages will be shown in the box? [09:03:51] for labels? [09:06:38] Yup and aliasses [09:07:10] We could make a more lightweight VIP's labels then [09:09:56] showing all of them in the table seems like it wouldn't be very easy to use (it would be huge), but I agree it needs a way to add stuff for a language not already there [09:10:19] aude: ? :D [09:11:04] nikki: I have something like the sitelinks widget in mind [09:11:06] I could imagine something similar to sitelinks, an empty row where instead of a label for the language, it has a search box (like how monolingual text has one) [09:11:08] heh :) [09:11:09] where you can add an additional language [09:11:16] haha [09:11:41] But a mass-editor for power users would be great. :) [09:16:16] Reedy: got time to look at https://gerrit.wikimedia.org/r/#/c/266932 again? :P [09:16:44] But I guess it's better to wait for more stable times. :) [12:26:42] https://www.wikidata.org/w/index.php?title=Q1126708&type=revision&diff=301853199&oldid=301639987 :) [12:30:19] addshore: I see you are signed up for a SWAT, would you also do https://gerrit.wikimedia.org/r/#/c/254645/3 ? [12:30:55] sure, is it ready to go as it is? [12:31:06] also, checking it works is just a case of using mobile search (from the app)? [12:32:07] it needs rebase [12:33:10] think i am going to move the setting [12:33:18] just so they are together [12:39:41] well, if it is ready for the SWAT I can do it ;) [12:40:12] addshore: it will be. to check: not from the app, just check if searching for a label on https://m.wikidata.org/wiki/Wikidata:Main_Page works instead of just for Qids [12:40:23] ahh okay! [12:40:26] * aude rebasing [12:49:47] addshore: ok it is ready now [12:50:54] already added it to swat [12:50:58] thx [12:51:08] * aude hopes 9 patches for swat is ok, when it's normally max of 8 [12:58:38] DanielK_WMDE: https://gerrit.wikimedia.org/r/#/c/267864/ [13:08:00] aude: what about 20 patches? [13:09:36] d3r1ck: probalby would make deployers mad :P [13:18:42] hmmmm, ok [13:55:01] jzerebecki: aude coolio! [14:33:09] JeroenDeDauw DanielK_WMDE should EntityDocument extend Comparable? [14:33:24] otherwise we don't have a guarantee that EntityDocuments have an equals method [14:43:51] Ehm, is it just me or is Wikidata fucked up [14:44:18] I get on m.wikidata.org for some reason and I see the desktop version [14:44:54] poor you :( [14:45:06] are you on mobile? [14:45:08] Weird fuckup. [14:45:09] Nope [14:45:12] because for me on desktop it works [14:49:36] benestar: i'd prefer to avoid that dependency on datavalues stuff on that level. e.g. i could imagine Comparable could get a "greater" method at some point. I#d suggest to just add equals() to the EntityDocument interface [14:50:48] DanielK_WMDE: hmm, we already use comparable in our data model, eg. on StatementList [14:51:02] Fingerprint as well [14:51:21] actually most of our data model classes already implement it [14:51:42] not happy about that [14:51:51] i actually don't see much use for that interface [14:51:58] when and where would you typhint against that? [14:52:23] so just add it to EntityDocument? [14:52:48] I think in the current state it would be easier to make EntityDocument extend it tbh [14:53:04] so we are at least consistent in what our data model looks like [14:53:06] how would that be easier? [14:53:13] less lines of code [14:53:23] more consistency to the other classes [14:53:39] *shrug* [14:53:41] however, the other classes aren't interfaces... [14:53:45] i don't have very strong feelings about that [14:53:54] i told you my preference [14:53:59] but you are looking at the code, i'm not [14:55:01] benestar: the things is: fingerprints and entities implementing the same interfaces that defined an equals method doesn't buy you anything. they are not interchangable, not comparable to each other. that's why i think that interface is pointles. [14:55:38] that's true [14:55:57] additionally, equals on Entities behaves differently (not counting the id) [14:56:25] true! that would actually be an argument to use a different name, even [14:56:57] well... maybe. probably not. oh, whatever ;) [15:01:48] DanielK_WMDE: should it then only accept EntityDocument params or still "mixed" as type? [15:02:10] benestar: only EntityDocument. To compare with anything else makes no sense. [15:02:16] another reason not to use the generic interface [15:02:47] maybe call it isEqual then? [15:04:02] well, perhaps an unnecessary breaking change [15:04:14] we might even consider the mixed vs EntityDocument change "too" breaking [15:05:24] JeroenDeDauw: you reading this? [15:07:52] benestar: addign the type hint would only break code that is already broken. [15:08:03] changing the name would be more disruptive [15:08:12] agre [15:08:13] also, isEquals isn't really clearer [15:08:32] if we use a different name, it should really have a different meaning [15:13:19] DanielK_WMDE: https://github.com/wmde/WikibaseDataModel/pull/615 [15:19:33] Why can't I mark this useless edit? https://www.wikidata.org/w/index.php?title=User_talk:Ash_Crow&action=compare-header-revisions&header_newRevision=sy0ajfjxiovcf631 [15:20:31] benestar: hm, $target instanceof self is a but scary. An equivalence relation should always be symmetrical. The sublcass relationship is anti-symmetrical. if $b is and instance of a subclass of $a, you could end up with $a->equals( $b ) but not $b->equals( $a ). [15:20:55] that's of course not new in your patch. just something to think about [15:20:59] DanielK_WMDE: that's an old problem... :P [15:21:10] equals just doesn't work that well in oop :S [15:21:16] it's actualyl a classical bug when implementing equals() [15:21:24] yes, indeed [15:21:50] you can enforce that types are exactly the same [15:22:05] however, then you violate the contract of replacebility for inheritance [15:22:19] * benestar wonders whether that word actually exists [15:23:28] benestar: the term for that is Liskov's Substitution Principle :) [15:23:50] thanks, always forget that name ;) [15:28:24] DanielK_WMDE: I think besides of tests I got completely rid of Entity in Wikibase.git this weekend and today [15:30:53] benestar: cool! I still hate the name "EntityDocument", but whatever ;) [15:46:18] benestar: got a specific question? I'm busy with non-wd stuff and not going to read the whole chat [16:28:35] DanielK_WMDE: can i move https://phabricator.wikimedia.org/T125516 to done? [16:30:53] aude: I'd like to document the proposed solution in more detail there. [16:31:31] hm, let me do that now... i was going to do some code experiments, but it's all convoluted api crap and will take a while. [16:34:25] benestar: the name of that method - equals - is actually not so good, as one would expect identity to be included for entities such as our entities (that wording...) [16:34:37] benestar: valueEquals or something would avoid this [16:34:41] more annoying break tho [16:34:58] benestar: thoughts? [16:36:48] yes, I wanted to avoid that breaking change because we then have to touch actually a lot of code... [16:37:38] aude: done [16:37:42] but I also see the point that it isn't actually the same equals that we have in other places [16:38:28] JeroenDeDauw, benestar: wasn't the original semantics to ignore the id only if it's null for one of the two entities to be compared? [16:38:45] ugh, that sounds even weirder :P [16:38:57] JeroenDeDauw, benestar: I can't think of a reason why we would consider two entities with two distinct IDs to be equal. [16:39:27] if you want to see if they have the same stuff... [16:39:28] * benestar doesn't either but he didn't write that code [16:39:55] the current approach allows you to do everything [16:40:00] you can just do an id check on top [16:40:09] including a weird id check [16:40:39] yes, I don't think we should actually change what equals does if there is no real reason for it [16:41:21] benestar: well, if you dont think we should rename the thing, then lets go ahead with this change [16:42:04] the rename would be a different issue because we don't change the semantics of the equals itself [16:42:13] we could have renamed it earlier and we can still rename it later [16:42:45] so yeah, lets go with the change as it is atm [16:43:26] benestar: if we want to rename it makes sense to do so now and not in another breaking change in another rel [16:43:43] that's true indeed [16:48:15] i don't think it should be renamed at this point [16:58:40] is there a plan about using more of wikidata into wikipedia infoboxes ? [17:00:07] rom1504: in general, yes, that is the idea. but that's up to the wikipedia comminuties [17:00:08] I'm seeing https://phabricator.wikimedia.org/T76229 and https://www.wikidata.org/wiki/Wikidata:Development_plan#Infobox_demos_.2B_documentation but what about actually changing a lot of wikipedia infoboxes ? [17:00:34] okay, maybe there would be some page talking about that in wikipedia then [17:00:41] we plan to make it easier to use wikidata on wikipedia, but what gets used, and when and where, is up to the local community [17:01:31] rom1504: i supposed there is :) [17:02:13] ah there's https://en.wikipedia.org/wiki/Wikipedia:Wikidata#Infoboxes_.28Phase_2.29 [17:02:25] it's still very theoritical though [17:03:30] to me it sounds like transforming infoboxes into wikidata thingy could be done semi-automatically for a lot of pages [17:03:40] I'm sure I'm not the first one to think about that [17:04:19] (of course it's up to mediawikis to decide what to do, but I think such a transformation could be interesting) [17:05:03] I'll ask #wikipedia I guess [17:08:49] ah this is interesting https://www.wikidata.org/wiki/Wikidata:WikiProject_Infoboxes [18:04:05] wohoo search works on mobile wikidata \o/ [18:04:06] aude --^ [18:10:19] benestar: in the browser, or in a native app? [18:17:57] benestar|cloud: \o/ [18:42:22] hi, is there some HOWTO to get started by editing wikidata from command line or from a programming language? [18:42:48] I've improved my setup, and editing trough the web interface is too slow, especially when the actions are repetitive and scriptable [18:42:51] you mean a guide to its API? [18:42:55] *is still too slow [18:43:03] ...or using bots like Pywikibot? [18:43:08] bots I guess [18:43:19] API would work too, if I can edit trough it and if it's easy enough [18:44:15] The idea is to do semi-automatic edits, not fully automatic. [18:44:35] I'm mapping hardware starting from SOCs [18:45:27] so for instance on the I.MX25 family of Freescale SOCs, most is duplicated between the I.MX25{1,3,5,7,8} [18:46:05] I also have some previous work I could insert into wikidata if I can script it [18:46:12] https://libreplanet.org/wiki/User:GNUtoo2/Hardware_freedom_documentation/SOCs [18:46:31] If it's manual, then I will probably loose interest [18:47:24] andre__: what's the difference between the API and bots in that context? [18:47:57] andre__: I never used bots, I only looked in some python thing (probably pywikibot) long time ago to create a GUI like commonist [18:48:34] or maybe it was pywikipedia or media I don't remember [18:53:16] https://www.mediawiki.org/wiki/Wikibase/API [18:53:25] https://www.mediawiki.org/wiki/Manual:Pywikibot [18:54:28] thanks [18:59:40] ok, I'll try pywikibot [19:00:09] I already looked at some SPARQL API, and this one seem like the GET API [19:00:27] as for pywikibot I guess I need to try it, I'll go trough the tutorial [19:00:53] I was wondering if there is anything specific to wikidata for pywikibot usage [19:01:04] but I guess not since you pointed me to the standard documentation. [19:01:23] I'll guess all will make more sense after I start using it. [21:13:02] aude: Lydia_WMDE: Guess what, I will be in Berlin next week [21:13:30] multichill: yay! :) wanna come over for a visit in the office? or beer in the evening? [21:13:51] Is that mutually exclusive? :P [21:14:05] weeeelllll... [21:14:06] :D [21:14:14] I was thinking of something like that. I will be at a conference during the days [21:14:19] ok [21:14:27] * Lydia_WMDE tries to finish weekly summary [21:15:35] Would be fun to hang out one of the nights. [21:15:42] cool [21:15:47] let me know when you know more [21:17:29] I miss Berlin. :( [21:17:35] multichill: \o/ [21:18:05] http://www.ciscolive.com/emea/details/week-at-a-glance/ is the program. Looks like Monday evening is free :-) [21:18:11] * aude is back in berlin and moved out of the hostel :) [21:18:33] We're a partner, but I'm not going to the partner day [21:18:56] monday is not a deployment day, so would be good [21:25:03] Great aude, you're no longer homeless ;-) [21:56:52] hoo: hey [21:57:26] hi d3r1ck :) [22:22:27] hoo: its been a while [22:22:37] Indeed [22:22:46] How are you experiments with Wikibase going? [22:22:53] i have been having a chat with Lydia, should i Bcc you in the email thread? [22:23:05] hoo: well i am trying and putting my all effort but its great [22:23:35] Sure, feel free to CC me [22:24:45] Did you manage to find an easy bug to work on for starters? [22:25:38] Hmmm, not yet :) [22:25:43] hoo: Check this out [22:25:45] https://phabricator.wikimedia.org/T118463 [22:25:54] will you be able to work with me on that? [22:26:19] Lydia is willing to co-mentor but she will need to talk with her team on it first, but i am just asking you :) [22:27:37] availability is probably an issue with me, as said earlier on [22:27:44] but I'm totally happy to mentor [22:27:58] but I can't guarantee swift responses all the time [22:28:57] hoo: what matters to me most hoo is not quick response but reliability [22:29:16] if you can always respond whether late or early, i am happy to take that [22:29:58] I just want someone to guide me as a mentor and i do all the work as planned. We will mostly communicate by email not even IRC [22:30:05] hoo: so is that OK by you [22:30:06] ? [22:31:07] I don't usually work everyday, so that might be a little tough for me… someone needs to fill these gaps [22:31:19] I can probably spend a bit of time daily, though [22:31:26] I still do email everyday [22:32:12] hmmmm, thats why co-mentores are available. And also that is why i choosed email for communication :) [22:32:41] so can i say i have a co-mentor and a mentor? [22:33:26] preliminary, yes, I guess. Still need Lydia to confirm the whole thing, but that sound sok to me [22:33:38] I think you will have to discuss with Lydia, so you both can confirm about it. She will probably bring the topic up on Wednesday like she said [22:33:58] hoo: Exactly, just like you said, you are perfectly correct. [22:34:40] If everything is fine an good by Wednesday, then i will continue work while waiting for the list of accepted orgs [22:35:29] hoo: also, you might want to post the project here if things a confirmed, https://phabricator.wikimedia.org/tag/google-summer-of-code-2016/ [22:35:45] that was created by Tony Thomas earliar today [22:37:00] hoo: Whats your email please so that i can Cc you in the thread? [22:37:16] marius.hoch [22:37:18] @ [22:37:20] wikimedia.de [22:37:58] Ok, you will receive a mail soon to confirm it reached you. [22:38:00] Thanks [22:40:45] hoo: did you receive any email? [22:40:58] I did, yes :) [22:41:14] hoo: Ok, we can continue from there since its the most reliable [22:51:48] it's a Lydia_WMDE on vacation? [22:52:03] she is AFAIK [22:52:49] hey hoo ;) [22:54:54] hi :)