IRC office hours/Office hours 2012-12-18a

From Meta, a Wikimedia project coordination wiki
Jump to: navigation, search

[17:31:12] <Lydia_WMDE> alright folks
[17:31:16] <Lydia_WMDE> let's get this started :)
[17:31:22] <Marco74> Hi
[17:31:24] <Lydia_WMDE> Denny will be a few mins late but we will start already
[17:31:30] <Lydia_WMDE> who's here for the office hour?
[17:31:36] <Marco74> Me
[17:31:40] <Lydia_WMDE> \o/
[17:32:11] <Marco74> We need to carefully look at the time today
[17:32:16] <DanielK_WMDE> Lydia_WMDE: i'm on and off. will pop in every now and then and try to answer questions
[17:32:18] <Marco74> as we need to switch languages
[17:32:23] <Lydia_WMDE> DanielK_WMDE: k - thx
[17:32:28] <Lydia_WMDE> Marco74: right
[17:32:31] <Lydia_WMDE> ok
[17:32:34] <Lydia_WMDE> let's start then
[17:32:38] <Lydia_WMDE> questions anyone?
[17:32:46] <Lydia_WMDE> comments?
[17:32:47] <Lydia_WMDE> wishes?
[17:32:50] <Lydia_WMDE> something else?
[17:33:13] <Marco74> As we have christmas in a few days ...
[17:33:17] <Marco74> ;-)
[17:33:24] <Lydia_WMDE> hehe
[17:34:21] -*- DanielK_WMDE bites his tongue not to start an onthological dispute
[17:34:33] <Lydia_WMDE> DanielK_WMDE: !!! :P
[17:35:32] <Lydia_WMDE> no questions? oh come on folks :D
[17:35:53] <Lydia_WMDE> don't be shy
[17:36:40] <Jarry1250> Hungarian Wikipedia?
[17:37:26] <Lydia_WMDE> Jarry1250: you mean a date? as it currently looks mid-january
[17:37:44] <Lydia_WMDE> provided the world doesn't end and all that
[17:38:04] <Jarry1250> Hmm, what's there to be done between then and now
[17:38:12] <Jarry1250> (apart from surviving the apocalypse)
[17:38:14] <Lydia_WMDE> there's christmas in between
[17:38:21] <Lydia_WMDE> where nothing happens unfortunately
[17:38:23] <Jarry1250> Oh, and Christmas :)
[17:38:26] <Lydia_WMDE> :D
[17:38:44] <Marco74> I assume that if it would be lauched now
[17:38:47] <Lydia_WMDE> it's simply that the foundation doesn't do deployments anymore before the end of the year
[17:39:15] <Marco74> and anything unexpected will happen nobody would fix it
[17:39:18] <Marco74> for weeks
[17:39:23] <Lydia_WMDE> right
[17:39:32] <Lydia_WMDE> which we need to avoid obviously
[17:39:38] <Jarry1250> Mmm, fair enough. But what's been done recently on it?
[17:39:48] <Lydia_WMDE> on what exactly?
[17:39:58] <Jarry1250> To work towards the deployment
[17:40:05] <Jarry1250> i.e. why not a deployment last week?
[17:40:07] <Jarry1250> (say)
[17:40:07] <Lydia_WMDE> the client code has been polished
[17:40:10] <Marco74> there is a discussion about representing values
[17:40:16] <Lydia_WMDE> and things added like being able to see repo changes on the client
[17:40:50] <Lydia_WMDE> in recent changes you can now see when something happens on wikidata that concerns the client wiki (hu:wp for example)
[17:41:01] <Lydia_WMDE> and we rolled this all out on test2
[17:41:13] <Lydia_WMDE> the tests there seem fine so we're ready for the next step
[17:41:18] <Jarry1250> Hmm, annoying to have missed the deployment window though, no?
[17:41:27] <Lydia_WMDE> quite so ;-)
[17:41:36] <Lydia_WMDE> but then again
[17:41:48] <Lydia_WMDE> it'll happen - a few days won't make that much of a difference
[17:42:09] <Lydia_WMDE> and in the end it is really better to do it when everyone is around to fix problems should they arrise
[17:42:15] <Lydia_WMDE> which we hope they won't obviously but...
[17:42:35] <DanielK_WMDE> we could discuss the Data Values thread that is currently going on wikidata-l
[17:42:46] <Denny_WMDE> sorry for being late
[17:42:50] <Jarry1250> So about another month then. *puts it in calendar*
[17:42:58] <Denny_WMDE> hello all
[17:43:04] -*- Lydia_WMDE waves @ Denny_WMDE
[17:43:20] <Lydia_WMDE> Jarry1250: yeah with the caveat that the world might end and all
[17:43:37] <Denny_WMDE> or that other issues arise
[17:44:04] <Jarry1250> Well yes, of course. Things can always arise.
[17:44:21] <Jarry1250> I had a noob question about the value representation question
[17:44:26] <Denny_WMDE> go ahead
[17:44:26] <Jarry1250> now I think about it.
[17:44:37] <Marco74> Hello Denny
[17:45:07] <Jarry1250> confidence interval is fine for scientific data, but what if you are uncertain but don't have an interval? Can you just not set confidence/min/max ?
[17:45:38] <Denny_WMDE> confidence min max is the same as an confidence interval
[17:46:00] <Jarry1250> Yes, so can just not set a confidence interval?
[17:46:10] <Denny_WMDE> yes
[17:46:36] <Jarry1250> Mmm, fair enough, it wouldn't imply confidence=1 then?
[17:46:39] <Denny_WMDE> my current thinking is: the systems assumes that all leading non-zero digits are significant if you dont say anything else
[17:47:05] <Denny_WMDE> and then set up automatically uncertainties accordingly
[17:47:27] <Denny_WMDE> you say, the height is 1200 meters
[17:47:37] <Denny_WMDE> and so it assumes, ah, it's 1200+-100 meters
[17:47:47] <Jarry1250> What if you measure it to 1243, but you don't know what your error range is
[17:48:04] <Jarry1250> *someone else reports it as
[17:48:49] <Jarry1250> I mean, you'd have thought that if they were reporting it as 1243 they mean +-1 as you suggest, but what if you don't *know* that?
[17:48:58] <Denny_WMDE> if a source gives you a value of 1243 meters, then you would take it at face value
[17:49:14] <Denny_WMDE> the system is no smarter than that
[17:49:47] <Denny_WMDE> maybe we will allow to choose undefined for these values
[17:50:07] <Denny_WMDE> sorry
[17:50:08] <Denny_WMDE> rephrase
[17:50:17] <Denny_WMDE> maybe we will allow to choose undefined for the uncertainties
[17:50:32] <Marco74> if somebody doesn't know it than he does not have a source
[17:50:44] <Denny_WMDE> but the UI will default to 1243+-1. you can click on detail and change that
[17:51:02] <Marco74> I would suggest N/A values
[17:51:09] <Marco74> for everything
[17:51:18] <Denny_WMDE> Marco74: most sources do not give uncertainties explicitly
[17:51:40] <Denny_WMDE> the problem with undefined or N/A is the unit transformation
[17:51:42] <Marco74> Ah, you mean that
[17:51:58] <Marco74> Hm.
[17:52:05] <Denny_WMDE> how can I translate 1243 meters into feet if i don't know the uncertainty?
[17:52:31] <Jarry1250> Well I suppose you assume +-1 for the uncertainty, but you don't display it to the user.
[17:52:38] <Jarry1250> (+- 0.5m)
[17:52:44] <Marco74> then the value will be around 1243
[17:52:51] <Marco74> which can be transcalced
[17:53:08] <Marco74> and you cannot transcalc the uncertainty
[17:53:15] <Marco74> because you do not have it
[17:53:42] <Marco74> then you have feet(1243) ± feet(n/a)
[17:53:54] <Denny_WMDE> yes, and that is pretty much how the system would work
[17:53:55] <Marco74> = feet(1243) ± n/a
[17:54:03] <Marco74> ah ok
[17:54:13] <Denny_WMDE> sorry, i meant to Jarry1250
[17:54:21] <Denny_WMDE> your solution would not work, Marco74
[17:54:46] <Marco74> y not?
[17:55:04] <Denny_WMDE> because feet(1243 m) = 4078.08 feet
[17:55:19] <Marco74> sorry I didn't transcalced it
[17:55:39] <Denny_WMDE> or more explicit, feet(1000 m) = 3280.84 feet
[17:55:47] <Denny_WMDE> if i say 1000 m
[17:55:51] <Marco74> so i meant 4078.08 ft. ± n/a ft
[17:56:07] <Denny_WMDE> but that would imply a precision that the original data didn't have
[17:56:18] <Denny_WMDE> suddenly you giving two numbers after the comma
[17:56:18] -*- Marco74 was to lazy to calculate
[17:56:42] <Denny_WMDE> we always must assume uncertainty
[17:56:56] <Denny_WMDE> or calculating to another unit will give results which do not make sense
[17:57:11] <Denny_WMDE> increasing accuracy due to transformation of units is a bad idea
[17:57:11] <Marco74> no the uncertainty was not available
[17:57:23] <Marco74> and this is what we are calculating with
[17:57:36] <Marco74> "not available" = n/a
[17:57:39] <Denny_WMDE> then you cannot translate units
[17:58:10] <Marco74> let's assume we have 1000 m ± 100 m
[17:58:30] <Marco74> that would be 1243 ft ± 124.3 ft
[17:59:06] <Denny_WMDE> what about 1200 ft +- 200ft?
[18:00:10] <Marco74> originating from a base value of 1200ft. which are 365.76 m
[18:00:22] <Marco74> we calculate the ± value
[18:00:48] <Marco74> which ist 200 ft. = 60.96 m
[18:01:01] <Lydia_WMDE> if anyone else has more questions for later send them my way and i'll queue them up
[18:01:25] <Marco74> We need at least to values
[18:01:36] <Marco74> wheater min/max pairs
[18:01:36] <Denny_WMDE> i consider translating 1200 ft +- 200ft as 365.76 m +- 60.96 m as a bad idea
[18:01:46] <Marco74> y
[18:01:47] <Marco74> ?
[18:01:58] <Denny_WMDE> because it looks much more accurate
[18:02:17] <Denny_WMDE> you have 4 significant digits
[18:02:19] <Denny_WMDE> where you had 1
[18:02:26] <Denny_WMDE> this does not make sense
[18:03:30] <Marco74> sorry I don't understand what you mean
[18:03:41] <Denny_WMDE> ok
[18:03:53] <Lydia_WMDE> Marco74: 1000 can be a pretty rough guess
[18:03:54] <Denny_WMDE> let us take the quantity value for the length of the equator
[18:03:57] <Lydia_WMDE> 1234.6 can't
[18:04:00] <Denny_WMDE> yes
[18:04:15] <Lydia_WMDE> and that's what Denny_WMDE is referring too
[18:04:15] <Denny_WMDE> if i say the equator is 40000 km
[18:04:23] <Lydia_WMDE> due to conversion you make it seem like it is very accurate
[18:04:26] <Lydia_WMDE> when it really isn't
[18:04:29] <Marco74> I just used the values been used before
[18:04:53] <Denny_WMDE> if i tell you the equator is 40000km
[18:05:03] <Denny_WMDE> or if i tell you the equator is 24854.823 miles
[18:05:14] <Denny_WMDE> do you really think this is the same statement?
[18:05:32] <Marco74> mathematically it's the same
[18:05:51] <Marco74> but I cannot imagine anything by 24854.823 miles
[18:05:58] <DanielK_WMDE> Denny_WMDE: if I understand correctly, you are assuming that Marco74 wants to deduce tha accuracy after convertig, instead of storing it directly after input
[18:06:12] <Denny_WMDE> well, a system that can tell that it is mathematically the same does not need that transformation anyway :)
[18:06:13] <DanielK_WMDE> DanielK_WMDE: but I think what he's really saying is that n/a means n/a and stays n/a.
[18:06:17] <DanielK_WMDE> "we don't know".
[18:06:26] <Lydia_WMDE> DanielK_WMDE: no it is what people will think when they see the more "detailed" number
[18:06:28] <Marco74> no, no
[18:06:45] <DanielK_WMDE> Lydia_WMDE: that's why the accuracy needs to be displayed.
[18:06:50] <Marco74> DanielK_WMDE: exacltly
[18:07:17] <Marco74> n/a is not available and always stays not available = n/a
[18:08:00] <DanielK_WMDE> Marco74, Denny_WMDE: i think *allowing* undefined or n/a for accuracy would be good, but the *default* should be what Denny suggested: go by the last non-zero digit.
[18:08:04] <Marco74> if somebody doesn't mention a value for tolerance then there is no value available
[18:08:41] <Marco74> DanielK_WMDE: that was the question?
[18:09:02] <DanielK_WMDE> Marco74: but in a statement like "it's a mile to teh supermarket" the accuracy is implicit. it's not two miles. but it may be 1.2 miles. the expected, real life, conversational accurary can reasonably be expected to be +/- 0.5 miles.
[18:09:29] <DanielK_WMDE> so, as a default, i'd say derive the accuracy upon input, and store it, so it can be converted along with the value.
[18:09:31] <Marco74> DanielK_WMDE: but it is not available
[18:09:45] <DanielK_WMDE> Marco74: yes it is. it's implicit in the notation of the value itself.
[18:10:15] <Marco74> DanielK_WMDE: yes because it is spoken and not really measured
[18:10:24] <Lydia_WMDE> fyi: we have 20 mins left
[18:10:28] <DanielK_WMDE> Marco74: if i write "the table is 4.3227 meters long", you can assume that i actually measured to sub-millimeter accuracy.
[18:10:41] <Jarry1250> I think there's a difference between conversion and how you display it. You shouldn't display it as 1 mile +- 0.5 miles if you don't know
[18:10:48] <Jarry1250> but you can use it for conversion.
[18:10:56] <Jarry1250> I don't agree DanielK, not for all values
[18:11:01] <DanielK_WMDE> Marco74: if someone states it, there's an implicite accuracy. ifd someone measures, there's a technical accuracy. so generally, there will be one
[18:11:04] <Jarry1250> There's a lot of false accuracy around
[18:11:31] <DanielK_WMDE> Jarry1250: true. so?
[18:11:38] <Marco74> Jarry1250: of course
[18:12:02] <DanielK_WMDE> Jarry1250: would it be better to treat all values as exact? or treat all values as completely unsure and thus unusable?
[18:12:17] <Jarry1250> Both.
[18:12:19] <Marco74>
[18:12:25] <DanielK_WMDE> i think both extremes are impractical. so, let's per default use the accuracy implicite in the notation
[18:12:29] <lazowik> Lydia_WMDE: You mean 20 min +- 10 min?
[18:12:42] <Jarry1250> Treat as unsure in the display, and the implicit (or explicit if present) accuracy for conversion
[18:12:47] <DanielK_WMDE> Jarry1250: both make the system a lot less useful
[18:12:57] <Lydia_WMDE> lazowik: no because the german one starts at :30
[18:13:32] <DanielK_WMDE> lazowik: hehe - technically, i guess Lydia_WMDE would have to specify the +/-1 explicitly :)
[18:13:38] <Lydia_WMDE> :P
[18:13:43] <lazowik> :D
[18:13:44] <Lydia_WMDE> +/- 0
[18:13:46] <Lydia_WMDE> ;-)
[18:13:58] <Marco74> who's clock
[18:13:59] <Marco74> ?
[18:14:07] <lazowik> freenode's
[18:14:08] <DanielK_WMDE> at what altitude?
[18:14:35] <Marco74> Danielk_WMDE: you mean lat/long?
[18:14:44] -*- Lydia_WMDE is part swiz and part german so very accurate in this ;-)
[18:14:57] <Lydia_WMDE> Marco74: mine obviously
[18:14:59] <Lydia_WMDE> ;-)
[18:15:07] <Lydia_WMDE> aaanyway - on with the actual discussion
[18:15:11] <Denny_WMDE> i don't think we can decide anything per office chat on this and the discussion should happen via Email and wiki
[18:15:21] <Lydia_WMDE> jep probably
[18:15:27] <Denny_WMDE> if there are questions or misunderstandings, that's what IRC is good at resolving
[18:15:41] <Lydia_WMDE> other topics?
[18:15:46] <Lydia_WMDE> questions? wishes? ...
[18:16:20] <Marco74> Yes:
[18:16:21] <DanielK_WMDE> Marco74: no, i mean the clock. higher altitude -> higher velocity -> greater time dillation -> slower clock ;)
[18:16:23] <Marco74> Licences
[18:17:07] <Lydia_WMDE> i don't think we can discuss more there before we heard back from the lawyers denny contacted
[18:17:16] <Denny_WMDE> depends
[18:17:19] <Denny_WMDE> what are the questions?
[18:17:38] <Lydia_WMDE> that would have been my next question ;-)
[18:17:44] <Lydia_WMDE> Marco74: ?
[18:17:51] <Marco74> DanielK_WMDE: that's true. ω is the same, v depends on the altitude you're correct
[18:17:55] <Marco74> Yes
[18:18:06] <Marco74> was just writing my previous answer
[18:18:21] <Denny_WMDE> the status is, we are using CC0 for the structured data and ccbysa for the unstructured text.
[18:18:22] <Marco74> do we have some answers concerning the questions about licences
[18:18:35] <Denny_WMDE> no, we didnt get answers yet from the lawyers
[18:18:47] <Denny_WMDE> it will probably take a few more weeks
[18:18:59] <Marco74> If the servers stood in Germany we could also use cc by sa germay 3.0
[18:19:07] <Marco74> sorry Germany
[18:19:21] <Denny_WMDE> the questions to the lawyers all concerned the topic of importing data from other databases to wikidata, and nothing else
[18:19:32] <Denny_WMDE> you mean for the text?
[18:19:41] <Marco74> for everything
[18:20:04] <Denny_WMDE> does ccbysa3 cover data?
[18:20:19] <Denny_WMDE> as far as i know it does not
[18:20:34] <Marco74>
[18:20:35] <Marco74> yes
[18:20:47] <Marco74> the Germany
[18:20:49] <Marco74> version
[18:20:56] <Marco74> the Austria version as well
[18:21:04] <Marco74> Soweit Datenbanken oder Zusammenstellungen von Daten einen immaterialgüterrechtlichen Schutz eigener Art genießen, unterfallen auch sie dem Begriff "Schutzgegenstand" im Sinne dieser Lizenz.
[18:21:19] <Marco74> found at the URL above
[18:21:30] <Denny_WMDE> so you saying CCBYSA-DE is not compatible to CCBYSA-unported?
[18:21:38] <Marco74> maybe?
[18:22:09] <Marco74> I first only looked fore ccby-unported, ccby-de and ccby-at
[18:22:24] <Marco74> ccby-unported lists many different types of works
[18:22:38] <Marco74> but doesn't discribe them
[18:23:06] <Marco74> ccby-at and ccby-de discribe them an only list them in parts
[18:23:30] <Denny_WMDE> well, anyway, i don't think ccbysa-de is an option. the servers are not in Germany
[18:23:48] <Marco74>
[18:23:58] <scott_wuas> Planning here to participate in the office hours in German ... My question in English has to do with when World University and School, (like Wikipedia with MIT OCW with free, online, CC, MIT-centric, university degrees planned, in many languages and countries) will be able to start moving our 537 wiki pages in a present wiki in English to Wikidata / Wikibase? Thanks.
[18:24:01] <Marco74> Anybody wrote on the mailinglist
[18:24:25] <Marco74> that also the nl version coveres databases
[18:24:42] <Marco74> may this could change it a little bit?
[18:25:25] <Lydia_WMDE> 5 mins
[18:25:46] <Denny_WMDE> no it doesnt
[18:25:54] <Denny_WMDE> because we are using the unported version
[18:26:01] <Denny_WMDE> that is what is stated on all wikipedias
[18:27:05] <Marco74> ok the unported says
[18:27:06] <Marco74> :
[18:27:14] <Lydia_WMDE> scott_wuas: hey :) as Denny_WMDE already explained last time Wikidata won't be a place for you to host a wiki and store wiki pages so i fear the answer to your question will be never
[18:27:26] <Marco74> "Work" means the literary and/or artistic work offered under the terms of this License including without limitation any production in the literary, scientific and artistic domain, whatever may be the mode or form of its expression including digital form, such as a book, pamphlet and other writing; a lecture, address, sermon or other work of the same nature;
[18:27:51] <Denny_WMDE> Marco74: are you having a question or do you want to have a discussion?
[18:28:32] <Marco74> sorry, my question was about the state with the lawyers
[18:28:44] <Marco74> so this is just for questions?
[18:29:00] <scott_wuas> Lydia: Will WUaS in an 'instance of Wikibase' relate to Wikidata?
[18:29:03] <Denny_WMDE> we are not here to discuss CC
[18:29:18] <Denny_WMDE> or to analyze it
[18:29:19] <Marco74> Denny_WMDE: ok
[18:29:29] <Denny_WMDE> especially since none of us, as far as i know, are lawyers
[18:29:37] <Denny_WMDE> so i am not sure what the point would be
[18:30:14] <Lydia_WMDE> scott_wuas: that depends on what you mean with that - i'm not sure
[18:30:30] <scott_wuas> THanks :) ...
[18:30:53] <Lydia_WMDE> alright - thanks everyone :)