User talk:Multichill/Archives/2015/July

From Wikidata
Jump to navigation Jump to search

Incorrect precision

In this edit the bot made an unverifiable edit, even allowing for the fact that Wikipedia is not a reliable source. The source, such as it is, states the latitude and longitude to the nearest arcsecond, so the quasi-verifiable precision is 1 arcsecond. But the bot gives a precision of 0.15065338562478°. This precision is neither verifiable nor understandable. Jc3s5h (talk) 14:46, 8 July 2015 (UTC)

Feel free to improve, this is a wiki you know. Multichill (talk) 16:45, 8 July 2015 (UTC)
I'm not a bot writer, and even if I were, fighting bots are not acceptable. Please explain the meaning of the precision value you gave, and your plan to correct the error. Jc3s5h (talk) 17:10, 8 July 2015 (UTC)
I don't consider this an error. I see you found phab:T89218. That's a more appropriate place to discus the implementation of coordinates. Multichill (talk) 19:29, 8 July 2015 (UTC)
You still have not explained how a reader can verify the precision is correct. If you are not willing to make your edits verifiable, I will have to ask to have the approval of the bot revoked. Jc3s5h (talk) 19:36, 8 July 2015 (UTC)
Be carefull with your wording, Jc3s5h... --Succu (talk) 20:01, 8 July 2015 (UTC)
I also don't see an error, so what needs to be explained, Jcessh? Edoderoo (talk) 20:42, 8 July 2015 (UTC)
If one says one imported a value from a source, and the added data consists of latitude, longitude, and precision, the it should be possible to look at the source and see how each of the three values was obtained from the source. The two successive edits I linked to at the beginning of this thread said the values were imported from English Wikipedia, and that the Latitude is 53.3961, the longitude is 9.77778, and the precision is 0.15065338562478 the mw:Wikibase/DataModel says all of these are in degrees.
So compare that to the source, the English Wikipeda. The coordinates are stated in "Infobox German location" as
|lat_deg = 53 |lat_min = 23 |lat_sec = 46
|lon_deg = 9 |lon_min = 46 |lon_sec = 40
The conversion from degrees-minutes-seconds is correct. But there is nothing in the article to indicate how precise the position is. In the absence of a specific statement about uncertainty, it is understood that competent science and technology writers will only include as many digits as are useful in presenting the value (this applies to final results, such as the values in the template, not necessarily intermediate results that are hidden from the reader). So the implied precision is a few arcminutes. Let us give the bot the benefit of the doubt and suppose the implied precision is +/- 9 arcseconds; if the precision were any worse than this, the 46 arcseconds should have been rounded to 50 arcseconds. Converted to decimal, that would be a precision of +/- 0.0025°. But the precision added by the bot, 0.1506533856247°, is 60 times worse. There is no other numerical information in the article that this value could have been derived from. In various places it has been suggested that the precision should relate to the geographic size of places like countries, municipalities, etc. But the article does not contain any such information, so even a proponent of that approach (which I am not) would not be able to understand where the number came from. Jc3s5h (talk) 21:19, 8 July 2015 (UTC)

Coors to humans

Hi, please learn your bot to not add Property:P625 to human items? I guess is not so hard. Thanks. --Jklamo (talk) 21:33, 7 July 2015 (UTC)

It's a relatively small edge case. You shouldn't remove the coordinates, but move them as a qualifier to the relevant claim. It's probably the final resting place of some person most of the time. Multichill (talk) 16:46, 8 July 2015 (UTC)
There were 400+ cases, i will not call that as small edge cases. I remove them all, as it is not clear, what relevant is relevant claim for them (and mostly it is not possible to find out it from wikipage "source"), it can be resting place, place of death, place of birth, place of residence... simply just coordinates for human is just half of the infomration. Please do not add them in future again (it is really not so hard). --Jklamo (talk) 22:57, 8 July 2015 (UTC)
Hmm, that's weird. If I recall correctly, the last time I imported coordinates I just got a couple of humans. Considering the fact that the bot imported over 20.000 items the error rate (2%) isn't that bad, but could be lower.
Did you remove them or are this coordinates still on the humans? I would like to look into this later. Multichill (talk) 06:58, 9 July 2015 (UTC)
I already removed them all. I remember the similar situation in past, when nobody was willing to sort out these because of difficulty of this. They are most of these, but not all, as there are also some other error already fixed (P625 -> P159 qualifier for company items, P625 in Q4167410 items - in fact items with mixed disambiguation and non-disambiguation iw links, that must be fixed manually). --Jklamo (talk) 12:01, 10 July 2015 (UTC)

Items for user subpages

These items are not notable per WD:N. A previous example is Q11317656.--GZWDer (talk) 18:16, 16 July 2015 (UTC)

"It fulfills some structural need, for example: it is needed to make statements made in other items more useful."? Multichill (talk) 18:18, 16 July 2015 (UTC)
However it's ill-defined. See Wikidata:Requests for comment/Wikidata:Notability overhaul for purposed clearify - This item is still non-notable per the future criteria.--GZWDer (talk) 18:24, 16 July 2015 (UTC)
Dammit GZWDer & Mbch331 can't you just be nice and not frustrate my work? These items fullfill a structural need, you broke User:NoclaimsBot. I'm going to undelete these items now. Multichill (talk) 21:14, 16 July 2015 (UTC)
If your bot is using it please do them at testwikidata: in case they will be deleted again.--GZWDer (talk) 06:42, 17 July 2015 (UTC)


Maybe its better to continue here. I need your help if you have the patient for that :) I have downloaded the kennethreitz-requests-v2.7.0-52-g9648301.tar. What do I have to do next? Xaris333 (talk) 19:09, 28 July 2015 (UTC)

Problem solved. Thnx! Xaris333 (talk) 02:34, 29 July 2015 (UTC)

Can I ask you some questions about the use of pywikibot? Xaris333 (talk) 17:30, 29 July 2015 (UTC)
This is just general python. Just run pip install requests on the commandline. That should install it. Multichill (talk) 21:11, 29 July 2015 (UTC)

I solved the problem yesterday. I can login now. But I have some questions about adding claims. I can add claims from infobox parameters (using harvest) and categories (using claimit). I wonder if there a way to add claims from a wikipedia table or from a list. Xaris333 (talk) 22:42, 29 July 2015 (UTC)