User talk:T.seppelt/Archive 1
This page is an archive. Please do not modify it. Use the current page, even to continue an old discussion. |
Welcome
Welcome to Wikidata, T.seppelt!
Wikidata is a free knowledge base that you can edit! It can be read and edited by humans and machines alike and you can go to any item page now and add to this ever-growing database!
Need some help getting started? Here are some pages you can familiarize yourself with:
- Introduction – An introduction to the project.
- Wikidata tours – Interactive tutorials to show you how Wikidata works.
- Community portal – The portal for community members.
- User options – including the 'Babel' extension, to set your language preferences.
- Contents – The main help page for editing and using the site.
- Project chat – Discussions about the project.
- Tools – A collection of user-developed tools to allow for easier completion of some tasks.
Please remember to sign your messages on talk pages by typing four tildes (~~~~); this will automatically insert your username and the date.
If you have any questions, don't hesitate to ask on Project chat. If you want to try out editing, you can use the sandbox to try. Once again, welcome, and I hope you quickly feel comfortable here, and become an active editor for Wikidata.
Best regards! Liuxinyu970226 (talk) 07:37, 30 December 2014 (UTC)
P1080 and P1441
Please, do not add P:1080 with works like in this. from narrative universe (P1080) should be used with fictional universes (see the constraints and the talk page). It's present in work (P1441) that should be used for indicate in which work the item appears. I corrected it but I would appreciate if you are more careful next time. Both P1080 and P1441 have 0 constraints violations in value now and I hope it can remain so. Many thanks! --Harmonia Amanda (talk) 10:06, 25 January 2015 (UTC)
- Thank you Harmonia Amanda, I didn't know this. I won't make this mistake anymore. Best wishes, -- T.seppelt (talk) 11:55, 25 January 2015 (UTC)
Error
Your bot is not marking its edits as bot edits, and thus is flooding recent changes. Can you fix that? George Edward C – Talk – Contributions 14:51, 25 January 2015 (UTC)
- I fixed it. Best wishes, -- T.seppelt (talk) 17:24, 25 January 2015 (UTC)
Refs error
Hi, i noticed, that your bot is sometimes adding just reference about importing from xxwiki without any real import. According Help:Sources statements that are only supported by "imported from (P143)" are not considered sourced statements, so adding just "imported from" "reference" without any real import does not make any sense. Please fix your bot to avoid these errors in future. Thanks. --Jklamo (talk) 22:09, 17 February 2015 (UTC)
- Hi Jklamo, I fixed this problem. Best wishes, --T.seppelt (talk) 12:06, 18 February 2015 (UTC)
- Thanks for fix. --Jklamo (talk) 13:03, 18 February 2015 (UTC)
Gender for duos
Hi T.seppelt,
At Q6433430, the bot added a gender. --- Jura 14:59, 21 February 2015 (UTC)
- Hi Jura,
- I used en:Category:American female singers as data set. It has unfortunately some mistakes. Excuse that please. --T.seppelt (talk) 15:03, 21 February 2015 (UTC)
- Don't worry, it's marginal compared to all the good additions it's doing. --- Jura 15:07, 21 February 2015 (UTC)
- It seems to happen again occasionally [1]. It might be worth checking the items for P31:Q5 before adding properties. --- Jura 19:22, 15 March 2015 (UTC)
- I check for P31:Q5 now. --T.seppelt (talk) 20:10, 15 March 2015 (UTC)
Edit summary viaf bot
Happy to see that you continued where I left off. Could you set a better edit summary like "adding NLP identifier A21374855 based on viaf"? Thank you, Multichill (talk) 22:31, 7 March 2015 (UTC)
- Yes, I will do this. You are welcome, --T.seppelt (talk) 22:33, 7 March 2015 (UTC)
- I was wondering, I assume you use viaf links from http://viaf.org/viaf/data/ (this file). Are you matching everything or are you restricting yourself to certain types? Multichill (talk) 15:42, 28 March 2015 (UTC)
- I only used certain types. Because of a licensing problem (#nonsensical source properties) I do not perform the VIAF thing any more. Your bot it doing something similar, isn't it? --T.seppelt (talk) 16:00, 28 March 2015 (UTC)
- Did you read meta:Wikilegal/Database Rights? Both Wikimedia and OCLC are based out of the US. Multichill (talk) 16:09, 28 March 2015 (UTC)
- Both are based in the US according to English Wikipedia. Gymel] has another opinion. According to [2] republishing of VIAF information is illegal if it doesn't happen "on the same terms and conditions as this License". Has CC0 same terms and conditions as ODC-By? --T.seppelt (talk) 16:22, 28 March 2015 (UTC)
- @Multichill: IANAL and thought that I stumbled upon some U.S. legal texts also concerned with "substantial parts of a database". Thus meta:Wikilegal/Database Rights might not show the full picture. That any single piece of data within VIAF is either CC0 or not copyrightable is out of the question. Thus either there are some rights hold by OCLC also in the U.S. or the whole thing of putting the data dumps under the ODC license is a case of copyfraud.
- Thursday OCLC resumed the discussion on en:WP announcing that they are about to switch from en:WP to Wikidata for their "Wikipedia" linking and ventilate a workflow for processing the constraint reports here. The original mapping VIAF to en:WP was performed by VIAF (i.e. OCLC) and my impression was that they kind of "donated" the result to en:WP (i.e. waiving their regular license. OTOH plain VIAF numbers are always fulfilling the ODC terms by design...). Thus getting an official statement from VIAF how they imagine (and tolerate) usage of VIAF data in Wikidata seems quite an option. -- Gymel (talk) 18:16, 28 March 2015 (UTC)
- Both are based in the US according to English Wikipedia. Gymel] has another opinion. According to [2] republishing of VIAF information is illegal if it doesn't happen "on the same terms and conditions as this License". Has CC0 same terms and conditions as ODC-By? --T.seppelt (talk) 16:22, 28 March 2015 (UTC)
- Did you read meta:Wikilegal/Database Rights? Both Wikimedia and OCLC are based out of the US. Multichill (talk) 16:09, 28 March 2015 (UTC)
- I only used certain types. Because of a licensing problem (#nonsensical source properties) I do not perform the VIAF thing any more. Your bot it doing something similar, isn't it? --T.seppelt (talk) 16:00, 28 March 2015 (UTC)
- I was wondering, I assume you use viaf links from http://viaf.org/viaf/data/ (this file). Are you matching everything or are you restricting yourself to certain types? Multichill (talk) 15:42, 28 March 2015 (UTC)
nonsensical source properties
please, sourcing as stated in (P248) e.g. LC numbers on VIAF entries like here does not make sense nor does it the other way round like there: VIAF asserts the pairwise identity of the persons identified by the corresponding authority file entries based on their contents and their utilization. Either the e.g. LCNAF record bears evidence for identification with the wikidata item or it does not, thus on intellectual inspection one could use the LCNAF record as source. For machine-derived identification with help of a VIAF listing stated in (P248) is much too strong... Likewise, even if the GND records in some representations contain VIAF numbers, they simply reflect (e.g. re-publish) the VIAF assertions (which may change every month) thus declaring GND as source stated in (P248) for VIAF number is circular at best. IMHO VIAF almost never can be used for sourcing wikidata claims, and if, then only with imported from Wikimedia project (P143), being the standard property for machine-derived data incorporation. And statements about VIAF numbers can only be backed by VIAF itself, i.e. tend to be tautological. -- Gymel (talk) 21:39, 8 March 2015 (UTC)
- I edited the sourcing sequence. Please look at this edit. I understand this problem. Circular referencing is even a big issue on the mostly human edited Wikipedia. (e.g. de:Wikipedia:Belege#Was sind zuverlässige Informationsquellen?). Adding VIAF identifiers based on GND entries or adding e.g. LCNAF identifiers based on VIAF entries is, in my eyes, quite reliable. A usual human editor would do it similarly. Even other bots have this task (example). How should I revise the program? Best regards, --T.seppelt (talk) 05:55, 9 March 2015 (UTC)
- Anyway, I interrupted the bot. It will not perform any edits until this discussion is closed. --T.seppelt (talk) 05:57, 9 March 2015 (UTC)
- Maybe I have fundamental objections against a) bots using stated in (P248) and/or b) that authority control usually implies "original work" (and therefore cannot be delegated to reliable, authoritative sources) but we shouldn't have to resolve that for the issue at hand;-).
- I think the situation is as follows: In the best case some authority number enters wikidata by extraction from some wikipedia article. Individual Wikipedias generally do not qualify as a source here, thus imported from Wikimedia project (P143) ist the appropriate qualifier in that situation, not stated in (P248). In other cases editors here enter authority numbers directly (hopefully aided by some gadget like User:Magnus Manske/authority control.js), in this case VIAF indeed might be considered a source for additional authority numbers (from personal experience I have to admit that I use the gadget's proposals quite blindly: VIAF does perform an excellent job with respect to clustering, but considering the association of individual authority records the matter is - by nature - less reliable, less stable and more volatile). Thus I pretty much can live with the current state of AC numbers individually imported from VIAF without any qualifiers.
- Now given some AC numbers in the wikidata item your bot tries to retrospectively qualify them: Via URLs like http://viaf.org/viaf/sourceID/GND|.... (or by using the justlinks data dumps) this is a simple lookup "AC number in, VIAF cluster Id out". Now we have a semantic pitfall, VIAF definitely is the (and the only authoritative) source for the statement "AC number x is associated with the VIAF Id y", but we are stating "the object represented by our wikidata item has the VIAF id y". This is something we (and not VIAF!) have derived from the association of "AC number x" with our item, and IMHO it cannot be sourced "stronger" than our original AC property (I consider stated in (P248) quite "strong" and especially much stronger than imported from Wikimedia project (P143)). [BTW: KrBot periodically imports AC statements from individual wikipedias and sources them with imported from Wikimedia project (P143): This is extremely helpful for editing because when some AC entry turns out to be wrong and is corrected here it can become quite challenging to find out whether some wikipedia should perform the same correction..]
- If your bot also imports additional AC numbers based on VIAF numbers (in the wikidata item), it should stop doing that: The ODC license demands "attribution" (which can easily be fulfilled by always providing a VIAF link). IMHO this license applies when VIAF is exploited systematically and at a large scale (there I see a difference between gadget-assisted power users which still import on the occasion of an individual record on one hand and bots striving to enrich everything on the other hand) and the trouble begins when the VIAF property should be accidentially deleted from the wikidata item or (more likely) someone re-using wikidata data utilizes isolated properties, e.g. by exporting a Wikidata-SUDOC mapping as with the Wikidata BEACON tool. To put it short: I consider the VIAF license not compatible with Wikidata's CC0 license when it comes to bot operations.
- On a more practical vein: My main objection is against your bot's usage of specifically stated in (P248) and I could live with imported from Wikimedia project (P143). -- Gymel (talk) 10:03, 9 March 2015 (UTC)
- Afterthought: Perhaps one could simply propose a new qualifiying property "derived from" or "inferred from"? -- Gymel (talk) 10:16, 9 March 2015 (UTC)
- I understand your opinion and the described issue. Do you agree with the following points? KasparBot does not exploit VIAF AC-data anymore (we don't need to discuss the difference between stated in (P248) and imported from Wikimedia project (P143), although I agree with you). I continue to import person information from GND (e.g. VIAF links) which are CC0. -- T.seppelt (talk) 14:08, 9 March 2015 (UTC)
- Probably it would be the best to add a hint to the property talk of stated in (P248) and imported from Wikimedia project (P143). -- T.seppelt (talk) 14:08, 9 March 2015 (UTC)
- I don't see licensing problems when importing VIAF numbers from whereever but would prefer the bot looking them up directly as http://viaf.org/viaf/sourceID/DNB%7Cgndnumber. Some GND recods also contain LCNAF numbers but usually these have been garbled so much on data entry that they are not worth the effort. I've come over some cases in the past where the VIAF cluster mixes several persons so I deliberately choose not to import the VIAF number into wikidata, but only selected individual AC numbers from the cluster, but these may be negligible and its clear to me that even the next human user will add the "missing" numbers from VIAF... Thus I'm alright with your bot adding VIAF numbers, provided it will use VIAF as source and attribute it to VIAF (not for licensing concerns but simply on methodical ground since I really do not consider GND records as appropriate source for "their" VIAF number). -- Gymel (talk) 17:48, 9 March 2015 (UTC)
- Okay. Fetching the AC links from http://viaf.org/viaf/sourceID/DNB%7Cgndnumber is from my point of view technically difficult. I will get the AC links by asking for justlinks.json. Then I will double-check whether the GND identifier from VIAF is the same as the GND identifier which is set at wikidata. If so, I add the other AC links from VIAF. If there is no GND identifier on wikidata the entry will be skipped. --T.seppelt (talk) 18:18, 9 March 2015 (UTC)
- Oh, I had understood you the other way round: Taking AC numbers (other than the VIAF number itself) not yet known here IMHO raises the license problem. Verifying or double checking numbers already known here and providing them with a source statement should be legitimate. Thus verifying and adding all VIAF numbers corresponding to any AC link already entered in a given wikidata item should be o.k. (well, maybe there is a conflict between verifying and adding?), but systematic and automated addition of non-VIAF identifiers by exploiting VIAF may constitute the crossing of a border: To my understanding the ODC license does not only apply to the VIAF datasets available for download but also to VIAF as a whole, e.g. as online database. (The ODC concerns "Database Rights: Database Rights only extend to the Extraction and Re-utilisation of the whole or a Substantial part of the Contents" - this is only granted when attribution to VIAF is maintained, for instance by providing the VIAF number in URL form. This is a non-restriction here but becomes an issue when data from here is reused - wikidata's license cannot and does not want to proliferate the ODC-BY terms to third parties). Wikidata currently stores more than 520.000 VIAF identifiers of almost 29M total VIAF clusters (February dumps) and has provisions to represent most of the AC identifiers covered by VIAF. Thus when the bot proceeds as announced, at the end Wikidata would mirror at least 2% of VIAF (that could well be more and will be more if one counts statements, not clusters). IANAL but "substantial parts" (of a database) seems to refer to quantitative and qualitative aspects and extends itself to repeated and systematic extraction of "insubstantial" parts (IMHO what the bot is aiming at). Therefore harvesting AC identifiers (other than VIAF identifiers themselves) from VIAF, dumps provided by VIAF, or VIAF-derived databases really should remain out of bounds for bots here! -- Gymel (talk) 21:09, 9 March 2015 (UTC)
- I agree with you. The bot will not perform any edits in such a way. Best regards, --T.seppelt (talk) 18:30, 10 March 2015 (UTC)
- Oh, I had understood you the other way round: Taking AC numbers (other than the VIAF number itself) not yet known here IMHO raises the license problem. Verifying or double checking numbers already known here and providing them with a source statement should be legitimate. Thus verifying and adding all VIAF numbers corresponding to any AC link already entered in a given wikidata item should be o.k. (well, maybe there is a conflict between verifying and adding?), but systematic and automated addition of non-VIAF identifiers by exploiting VIAF may constitute the crossing of a border: To my understanding the ODC license does not only apply to the VIAF datasets available for download but also to VIAF as a whole, e.g. as online database. (The ODC concerns "Database Rights: Database Rights only extend to the Extraction and Re-utilisation of the whole or a Substantial part of the Contents" - this is only granted when attribution to VIAF is maintained, for instance by providing the VIAF number in URL form. This is a non-restriction here but becomes an issue when data from here is reused - wikidata's license cannot and does not want to proliferate the ODC-BY terms to third parties). Wikidata currently stores more than 520.000 VIAF identifiers of almost 29M total VIAF clusters (February dumps) and has provisions to represent most of the AC identifiers covered by VIAF. Thus when the bot proceeds as announced, at the end Wikidata would mirror at least 2% of VIAF (that could well be more and will be more if one counts statements, not clusters). IANAL but "substantial parts" (of a database) seems to refer to quantitative and qualitative aspects and extends itself to repeated and systematic extraction of "insubstantial" parts (IMHO what the bot is aiming at). Therefore harvesting AC identifiers (other than VIAF identifiers themselves) from VIAF, dumps provided by VIAF, or VIAF-derived databases really should remain out of bounds for bots here! -- Gymel (talk) 21:09, 9 March 2015 (UTC)
- Okay. Fetching the AC links from http://viaf.org/viaf/sourceID/DNB%7Cgndnumber is from my point of view technically difficult. I will get the AC links by asking for justlinks.json. Then I will double-check whether the GND identifier from VIAF is the same as the GND identifier which is set at wikidata. If so, I add the other AC links from VIAF. If there is no GND identifier on wikidata the entry will be skipped. --T.seppelt (talk) 18:18, 9 March 2015 (UTC)
- I don't see licensing problems when importing VIAF numbers from whereever but would prefer the bot looking them up directly as http://viaf.org/viaf/sourceID/DNB%7Cgndnumber. Some GND recods also contain LCNAF numbers but usually these have been garbled so much on data entry that they are not worth the effort. I've come over some cases in the past where the VIAF cluster mixes several persons so I deliberately choose not to import the VIAF number into wikidata, but only selected individual AC numbers from the cluster, but these may be negligible and its clear to me that even the next human user will add the "missing" numbers from VIAF... Thus I'm alright with your bot adding VIAF numbers, provided it will use VIAF as source and attribute it to VIAF (not for licensing concerns but simply on methodical ground since I really do not consider GND records as appropriate source for "their" VIAF number). -- Gymel (talk) 17:48, 9 March 2015 (UTC)
Danke für die Ergänzungen
Könntest Du aber Deinem Bot aber bitte noch beibringen bei Geburtsjahr ca. und dergleichen in der GND noch den entsprechenden Qualifikator (sourcing circumstances (P1480)) zu setzen? Vgl. http://d-nb.info/gnd/101669968 und und diesen Difflink. --HHill (talk) 08:42, 9 March 2015 (UTC)
- Ja, kümmere mich darum. Viele Grüße, --T.seppelt (talk) 14:31, 9 March 2015 (UTC)
- Der Bot fügt jetzt auch solche Qualifikatoren hinzu (Beispiel). In der GND sind ungenaue Daten sehr uneinheitlich gespeichert. Deswegen werden manche Angaben nicht erkannt. Für die meisten Fälle sollte es aber funktionieren. --T.seppelt (talk) 18:08, 9 March 2015 (UTC)
Invalid LAC codes
Hello, please see this edit. I do not find 000001938 LAC code on https://viaf.org/viaf/85123459/ page. Looks like bot`s bug. — Ivan A. Krestinin (talk) 19:49, 9 March 2015 (UTC)
- Hello, for this entity VLACC-000001938 and LAC-0010C8143 were mixed up at VIAF. Both are set as value for LAC at https://viaf.org/viaf/85123459/justlinks.json. Anyway, the bot isn't performing this task at the moment because of an ongoing discussion (see above). I will fix this problem by using regular expressions more intensively. Thank you, --T.seppelt (talk) 19:55, 9 March 2015 (UTC)
Koordinaten
Hallo. Sei vorsichtig beim Hinzufügen von coordinate location (P625). Koordinaten für Satelliten [3] und Personen [4] geben im Allgemeinen keinen Sinn. --Pasleim (talk) 12:40, 13 March 2015 (UTC)
- Hallo. Ich habe mit dem Hinzufügen von Koordinaten generell aufgehört. --T.seppelt (talk) 12:48, 13 March 2015 (UTC)
- Same for companies. For them coordinate location (P625) can be used as qualifier (mostly for) headquarters location (P159). It is true that sometime headquarters location (P159) is not present, but that is not reason to import coordinate location (P625) incorrectly, but more likely reason for import (first) headquarters location (P159) from company article infoboxes (from de Infobox Unternehmen - Sitz or from en Infobox company - location). --Jklamo (talk) 18:53, 14 March 2015 (UTC)
- And if there is already coordinate location (P625) as qualifier for headquarters location (P159), there is no need to add coordinate location (P625) directly again. --Jklamo (talk) 19:01, 14 March 2015 (UTC)
- And same for disambiguation pages. --Jklamo (talk) 19:07, 14 March 2015 (UTC)
- As I mentioned before, I wont add coordinate location (P625)-claims any more. Future programs will always include a P31-doublecheck. Best regards, --T.seppelt (talk) 05:43, 16 March 2015 (UTC)
- Same for companies. For them coordinate location (P625) can be used as qualifier (mostly for) headquarters location (P159). It is true that sometime headquarters location (P159) is not present, but that is not reason to import coordinate location (P625) incorrectly, but more likely reason for import (first) headquarters location (P159) from company article infoboxes (from de Infobox Unternehmen - Sitz or from en Infobox company - location). --Jklamo (talk) 18:53, 14 March 2015 (UTC)
P512 and Q4618975
Hi there, this edit by KasparBot to Hanns Jürgen Küsters (Q1576469) set P512 to Doctor (Q4618975), but Q4618975 cannot be used with P512 because it is not a subclass of academic degree (Q189533) per Property_talk:P512. Maybe it should be doctorate (Q849697)? If Q4618975 is the best choice, then it should be connected to Q189533. These are the subclasses of 189533 right now: http://tools.wmflabs.org/wikidata-todo/tree.html?q=189533&rp=279&lang=de Thanks --Haplology (talk) 03:50, 14 March 2015 (UTC)
- Hi Haplology, I thought I checked it for constraint violation... Anyway, I now add doctorate (Q849697) with this program. I run a Widar-task right now to replace all academic degree (P512)Doctor (Q4618975) with academic degree (P512)doctorate (Q849697). Best regards, --T.seppelt (talk) 06:48, 14 March 2015 (UTC)
- There seems to happen much editing back and forth at the moment - in some languages it seems easy to differentiate between Doctor as title and Doctor as degree, in others this is a quite artificial distinction. However it seems clear to me that Wikidata does make that distinction. -- Gymel (talk) 17:26, 16 March 2015 (UTC)
von
Thanks for all the editions you recently made.
BTW at [5], somehow part of the name got lost. --- Jura 13:07, 14 March 2015 (UTC)
- I solved this problem. Thank you, --T.seppelt (talk) 14:46, 14 March 2015 (UTC)
Old data?
Hi, here your bot added a VIAF number "based on GND", however the GND ID (P227) in the item (101722850, I presume your bot based its action on that) had been deleted almost five days ago (here on wikidata and on de:wp)! Secondly, wasn't the consensus from our discussion some days ago, to not declare "GND" as source for VIAF numbers (they aren't even "in" the GND in the sense of "stored", DNB just adds them specifically to the RDF/XML representation of the records and chances are about 100% that they use the monthly VIAF dumps as base for that)? -- Gymel (talk) 16:50, 20 March 2015 (UTC)
- I solved both problems. The VIAF number issue results from a misunderstanding on my side. It will not occur any more. Best regards, --T.seppelt (talk) 09:53, 21 March 2015 (UTC)
Malformed IMDb IDs
Hello. Please make sure that KasparBot doesn't add malformed IMDb IDs to items. You can use this simple regex:
(tt|nm|ch|co|ev)[0-9]{7}
Thanks. Mushroom (talk) 01:29, 27 March 2015 (UTC)
- Hello. I am sorry. I fixed it. Thank you, --T.seppelt (talk) 05:42, 27 March 2015 (UTC)
Wrong description
Hello, Your bot added ml description to Q9045 which was in English language. There seem to be a mistake. I undid it for now because ml is a non latin script language.--Vyom25 (talk) 10:00, 29 March 2015 (UTC)
- Hello Vyom25, I used information from ml:Template:Persondata. I am sorry for the mistakes, --T.seppelt (talk) 13:55, 29 March 2015 (UTC)
- Oh okay, but be careful.--Vyom25 (talk) 05:26, 30 March 2015 (UTC)
- Also here. If you're using Java, something like langdetect could help you skip descriptions which surely aren't in the desired language. Thanks, --Ricordisamoa 07:18, 31 March 2015 (UTC)
- Thank you. It looks interesting, --T.seppelt (talk) 20:14, 31 March 2015 (UTC)
- I reverted quite a few of those as well. Could you maybe look into reverting all your edits that got the tag to indicate that a label was added in a script that is likely wrong? This one: https://www.wikidata.org/w/index.php?title=Special:RecentChanges&tagfilter=adding+non-latin+script+language+description+in+latin+script --LydiaPintscher (talk) 08:37, 2 April 2015 (UTC)
- I am working on a solution. It would be quite easier with the rollback privilege. --T.seppelt (talk) 14:04, 2 April 2015 (UTC)
- P.S.: it is working. The bot deletes the malformed description right now. Regards, --T.seppelt (talk) 14:15, 2 April 2015 (UTC)
- Awesome. Thank you :) --LydiaPintscher (talk) 15:14, 2 April 2015 (UTC)
Mojibake
Hi T.seppelt, at Q179065 some characters didn't quite turn out. --- Jura 15:41, 31 March 2015 (UTC)
- Hi Jura1, I am going fix this bug. --T.seppelt (talk) 20:14, 31 March 2015 (UTC)
- I did some test edits. The problem doesn't occurs any more. I think the java encoding defaults changed on tool server... --T.seppelt (talk) 20:29, 31 March 2015 (UTC)
- I tried to find a few P31:Q5 with "?" in the en label. Here are all labels of these items: http://quarry.wmflabs.org/query/2889 . Some might be ok.--- Jura 00:03, 1 April 2015 (UTC)
- Your query was quiet helpful. I was able to edit 136 entities with encoding problems. --T.seppelt (talk) 13:41, 1 April 2015 (UTC)
- I refreshed the query and improved it slightly. Everything seems fine now. --- Jura 08:51, 2 April 2015 (UTC)
- Your query was quiet helpful. I was able to edit 136 entities with encoding problems. --T.seppelt (talk) 13:41, 1 April 2015 (UTC)
- I tried to find a few P31:Q5 with "?" in the en label. Here are all labels of these items: http://quarry.wmflabs.org/query/2889 . Some might be ok.--- Jura 00:03, 1 April 2015 (UTC)
VIAF to lists and other
Hello, please see [6]. Could you add to your bot check for P31 = Q5? Authority control template is used too widely in Wikipedia unfortunately. This causes type errors and duplicate values in Wikidata. — Ivan A. Krestinin (talk) 13:51, 6 April 2015 (UTC)
- Hello Ivan, I check now if an item has instance of (P31) Wikimedia disambiguation page (Q4167410), Wikimedia category (Q4167836), Wikimedia list article (Q13406463) or Wikimedia article page (Q15138389). In that case it will be skipped. instance of (P31)human (Q5) doesn't fit with companies, places or ideas. Best wishes, --T.seppelt (talk) 17:05, 6 April 2015 (UTC)
- Thank you, ok, lets try to create more general criteria. More cases: [7], [8], [9], [10], [11]. — Ivan A. Krestinin (talk) 18:19, 6 April 2015 (UTC)
- @Ivan A. Krestinin: Obviously the errors were introduced in en:WP and had to be corrected there (which I just did). At least two of your examples could have carried authority control numbers in principle, the numbers imported where "just" plain wrong. Importing stuff like this to wikidata can help cleaning up the individual wikipedias, this is IMHO valuable enough to tolerate errors (temporarily) here. -- Gymel (talk) 20:24, 6 April 2015 (UTC)
- @Gymel: I requested the bot flag on enwiki to remove redundant information from the template. Any support would be great --T.seppelt (talk) 13:03, 8 April 2015 (UTC)
- @Ivan A. Krestinin: Obviously the errors were introduced in en:WP and had to be corrected there (which I just did). At least two of your examples could have carried authority control numbers in principle, the numbers imported where "just" plain wrong. Importing stuff like this to wikidata can help cleaning up the individual wikipedias, this is IMHO valuable enough to tolerate errors (temporarily) here. -- Gymel (talk) 20:24, 6 April 2015 (UTC)
- Thank you, ok, lets try to create more general criteria. More cases: [7], [8], [9], [10], [11]. — Ivan A. Krestinin (talk) 18:19, 6 April 2015 (UTC)
- Similar concern, this is incorrect, at Seth and Mary Eastman (Q15975279). The WP page describes a group of people, while the VIAF is only for one person (that happens to be Seth Eastman (Q931422)). Please exclude items derived from group of humans (Q16334295). -- LaddΩ chat ;) 19:38, 8 April 2015 (UTC)
- Again, I fixed the issue on en:Seth and Mary Eastman.
- Personally, I expect the bot to import that kind of crap from en:WP: For the Authority Control template the rules on en:WP are in alignment with the practice here at Wikidata, thus these errors should not be there. But they of course are and can obviously more easily be spotted here than there. I do see a conflict in policy between Wikidata "protecting" itself against erroneous data vs. Wikidata providing auxiliary services for the individual wikipedias, but maybe this is not the right place to discuss an issue that general. -- Gymel (talk) 21:29, 8 April 2015 (UTC)
- I see how you fixed it. Thanks, will do the same if I encounter such issues again. -- LaddΩ chat ;) 01:24, 10 April 2015 (UTC)
KasparBot
[Copy of what I wrote on your en.Wikipedia user page, in case you don't see it there]
I note from en:Wikipedia:Bots/Requests for approval/KasparBot that "If Wikidata has conflicting information the Article will be skipped". Instead of simply skipping the article, could the bot please also write an entry to a log, which we can then manually check? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:26, 26 April 2015 (UTC)
Marguerite MacIntyre
Special:Diff/214076934. Look at the history. --Marek Koudelka (talk) 07:49, 3 May 2015 (UTC)
- This isn't a bug. The edits were based on GND ([12]). I removed the GND claim. It should be right now. Regards, --T.seppelt (talk) 13:46, 3 May 2015 (UTC)
Check the history. Difficult one to solve. I guess we need to modify the authority control template on enwp to be able to give another Wikidata item id for which is valid. That would solve this case and the case with the authority control template on redirect articles. Multichill (talk) 18:49, 4 May 2015 (UTC)
- This would be the best option. But I think the wikipedia software isn't able to retrieve information from other wikidata items yet. Until such an improvement the current situation solves the problem. No LCNAF identifier on murder of Kitty Genovese (Q18341392), but on Kitty Genovese (Q238128). A big problem is that many wikipedians don't understand that wikidata items don't describe wikipedia articles but entities... Regards, --T.seppelt (talk) 19:39, 4 May 2015 (UTC)
- We're almost there, see phab:T49930. Usage tracking is already deployed on two wiki's so we're getting somewhere. Multichill (talk) 21:31, 4 May 2015 (UTC)
- Great! --T.seppelt (talk) 08:34, 5 May 2015 (UTC)
- We're almost there, see phab:T49930. Usage tracking is already deployed on two wiki's so we're getting somewhere. Multichill (talk) 21:31, 4 May 2015 (UTC)
"imported from" reference
Hello,
I noticed your bot added a imported from Wikimedia project (P143) English Wikipedia (Q328) reference instead of French Wikipedia (Q8447) for the authority information imported from the french Wikipedia. Example here. I'm not sure it's a very big deal, but I guess you should correct it.
Regards,
El pitareio (talk) 22:14, 26 May 2015 (UTC)
- Hello El pitareio,
- I noticed that problem and fixed it ([13], line 22). Maybe I didn't corrected all references on Wikidata. Thank you for doing that. Regards, -- T.seppelt (talk) 05:11, 27 May 2015 (UTC)
Koordinaten in Listen
Hi, solche Bearbeitungen sind natürlich unsinnig, eine Liste hat keine Koordinaten. ;-) Wär cool wenn du das beachten könntest und den bot fixen würdest. Viele Grüße -- Bene* talk 17:29, 6 June 2015 (UTC)
- Ja. Wenn ich das Programm irgendwann wieder benutze werde ich das beachten. Danke für den Hinweis --T.seppelt (talk) 21:42, 6 June 2015 (UTC)
Hello, I would like to inform that your bot request has been approved in Korean wikipedia. Please use wisely. Many thanks,--DangSunM (talk) 13:06, 1 July 2015 (UTC)
- Thanks -- T.seppelt (talk) 17:09, 3 July 2015 (UTC)
Authority control validator for Wikisource too
Hi,
I just discovered Authority control validator and it seems to be an fantastic tool. I just have request : is it possible to add the wikisources in it ? (on Template:Authority control (Q3907614) you can see the templates name on different wikisources).
Cdlt, VIGNERON (talk) 16:39, 23 July 2015 (UTC)
- @VIGNERON: Yes, I am working on it. Please notice that frwikisource decommissioned the local parameters (step 4 according to Wikidata:WikiProject Authority control/Status). For this reason we don't need to track differences for that project. Regards, -- T.seppelt (talk) 18:45, 23 July 2015 (UTC)
- Wonderful ! Yeah, I know that frws is a bit ahead but it can still benefit from verifications ont the others wikisources. Cdlt, VIGNERON (talk) 06:21, 24 July 2015 (UTC)
- @VIGNERON: please have another look at Wikidata:WikiProject Authority control/Status. bnwikisource, elwikisource, frwikisource and ukwikisource have completed step 4. plwikisource has to complete step 1 before tracking is useful. ruwikisource has a very different template which I have to inspect first. But you can find the errors for ptwikisource and enwikisource now on [14]. Be aware that the bot needs some time to cache all information. Regards, -- T.seppelt (talk) 06:59, 24 July 2015 (UTC)
- Wonderful ! Yeah, I know that frws is a bit ahead but it can still benefit from verifications ont the others wikisources. Cdlt, VIGNERON (talk) 06:21, 24 July 2015 (UTC)
Commons-Kategorien
Hi T.seppelt, ich sehe gerade auf meiner Beobachtungsliste, dass du ganz viele Site-Verlinkungen zu Commons-Kategorien entfernst, z.B. diese hier. Meines Wissen nach war es bisher Usus, dass diese Einträge OK waren, solange es keine konkurrierenden Kategorie-Datensätze, Commons-Galerieseiten o.ä. gab. Hat sich hieran etwas geändert? Raymond (talk) 15:03, 12 August 2015 (UTC)
- @Raymond: es geht bei den vielen Änderungen gerade um die Einträge, die einen Hauptkategorie zum Artikel-Claim haben. Der Commonssitelink wird entfernt und bei der zugehörigen Hauptkategorie hinzugefügt. Das sollte dem gängigen Vorgehen entsprechen. Viele Grüße, --T.seppelt (talk) 16:54, 12 August 2015 (UTC)
- Would probably be good to mention that in the (removing) edit summary: "<removing yadaa> moving the commons sitelink to <other item> based on topic's main category" and similar message when you add it. That makes mistakes like Category:Newspaper companies of Germany (Q9203855) easier to debug. Multichill (talk) 18:41, 12 August 2015 (UTC)
- I added a edit summary for both types of edits: removing and adding. -- T.seppelt (talk) 19:19, 12 August 2015 (UTC)
- Great! That should prevent more confusion :-) Multichill (talk) 07:58, 14 August 2015 (UTC)
- I added a edit summary for both types of edits: removing and adding. -- T.seppelt (talk) 19:19, 12 August 2015 (UTC)
- Would probably be good to mention that in the (removing) edit summary: "<removing yadaa> moving the commons sitelink to <other item> based on topic's main category" and similar message when you add it. That makes mistakes like Category:Newspaper companies of Germany (Q9203855) easier to debug. Multichill (talk) 18:41, 12 August 2015 (UTC)
- I have found and fixed more. Hier is some (not all):
- Can you also check category's main topic (P301) in main category? Most is missing P301, but must items on the liste is where category use in wrong way. --Steenth (talk) 11:01, 16 August 2015 (UTC)
- @Steenth: Yes, I can do this additional check. But it is good to see that the bot's edits helped to uncover that many inconsistencies. Thank you for reviewing, -- T.seppelt (talk) 08:09, 17 August 2015 (UTC)
Commons links
Hello, your bot removed link to ruwiki from page commons:Category:Titan (rocket). This happens die to [15] and [16] edits. Earlier sidebar on this page allowed one click navigation to ruwiki. Could you restore the link using some way? — Ivan A. Krestinin (talk) 20:12, 12 August 2015 (UTC)
- @Ivan A. Krestinin: commons categories should be connected to instance of (P31)Wikimedia category (Q4167836) entries if they exist. The bot is looking for items with Commons category (P373) or commons sitelinks and moves the sitelink to the item which is associated using a topic's main category (P910) claim. I don't know how to receive such links in commons (connected WD item → category's main topic (P301) → commons sitelink ??) but connecting commons categories (category namespace) with Wikipedia articles (main namespace) is in my eyes illogical. Regards, -- T.seppelt (talk) 21:28, 13 August 2015 (UTC)
- Wikidata item can connect different namespaces to one item. It depends on target project structure. For example Wikisource connects Author namespace to Wikipedia main namespace, Wikinews connects Category namespace to Wikipedia main namespace too. There was many discussions about Commons, but good solution was not found in it... About current issue: primary sitelinks usage for our users is navigation. Now navigation is broken. There are many ways to restore it: 1. revert your bot edits; 2. add local interwiki links to page commons:Category:Titan (rocket); 3. create template/module that will query sitilinks throw category's main topic (P301); 4. create script/gadget for this purpose; 5. some other way. — Ivan A. Krestinin (talk) 11:46, 15 August 2015 (UTC)
- @Ivan A. Krestinin: connecting different namespaces in different projects isn't a problem. But in this case Wikidata's aim is to represent entities (person, buildings etc. and categories). A category can never be the same entity as e.g. a person. As a tradeoff we connect Commons categories with non-category Wikidata entities when no corresponding category Wikidata entity is available. My bot edited Wikidata entities according to that rule. But back to the current situation: I would prefer solution 3 and until this is implemented solution 2. Regards, -- T.seppelt (talk) 08:15, 17 August 2015 (UTC)
- Wikidata item can connect different namespaces to one item. It depends on target project structure. For example Wikisource connects Author namespace to Wikipedia main namespace, Wikinews connects Category namespace to Wikipedia main namespace too. There was many discussions about Commons, but good solution was not found in it... About current issue: primary sitelinks usage for our users is navigation. Now navigation is broken. There are many ways to restore it: 1. revert your bot edits; 2. add local interwiki links to page commons:Category:Titan (rocket); 3. create template/module that will query sitilinks throw category's main topic (P301); 4. create script/gadget for this purpose; 5. some other way. — Ivan A. Krestinin (talk) 11:46, 15 August 2015 (UTC)
Incorrectly
1, 2 (Category:Honored Master of Sports…, but not Masters).
- @NBS: I don't get the divergence in content. Can you explain the difference between this entities? Thanks, -- T.seppelt (talk) 08:19, 17 August 2015 (UTC)
sv:Minnesmärke
On sv:Minnesmärke kasparbot/ac.php detects "unknown alias embedded or recursive transclusion". THe template used on that page is "auktoritetsdata". It is the normal template which sv.wp uses, only wiht a small letter 'a', not sure if that is the cause. (t) Josve05a (c) 04:02, 21 August 2015 (UTC)
- @Josve05a: It is probably caused by the small letter. I am going to fix this bug. Thanky, -- T.seppelt (talk) 18:20, 29 August 2015 (UTC)
Your bot is adding lots of good info
but (sorry there is always a but) could it add references too? even if it is just a reference URL (P854). Please? Joe Filceolaire (talk) 20:57, 22 August 2015 (UTC)
- @Filceolaire: My bot usually adds references. Can you provide any examples? Kind regards, -- T.seppelt (talk) 18:19, 29 August 2015 (UTC)
depositary (P2058) is ready. --Tobias1984 (talk) 11:23, 13 September 2015 (UTC)
- @Tobias1984: I began to import claims from enwiki (see Special:Contributions/KasparBot). Warm regards, -- T.seppelt (talk) 11:44, 13 September 2015 (UTC)
- Great! Indeed a much needed property. --Tobias1984 (talk) 11:45, 13 September 2015 (UTC)
- An error occurred. I fixed the problem now. Sorry ;-) -- T.seppelt (talk) 12:08, 13 September 2015 (UTC)
- Great! Indeed a much needed property. --Tobias1984 (talk) 11:45, 13 September 2015 (UTC)
Nicht-individualisierte "Normdaten"
Hallo T.seppelt, KasparBot importiert auch nicht-individualisierte "Normdaten", wie beispielsweise den Platzhalter für den Namen "Burney, Charles Fox" (GND, Typ n, siehe: Property talk:P227#Usage note). Ist es möglich, diese Fehleinträge in einer Liste auszugeben, dass sie möglichst bald korrigiert werden können? (Im vorliegenden Fall bei Wikisource.) Wo als Quelle VIAF angegeben ist, wäre es hilfreich, wenn ein Bot die Tns automatisch löscht. --Kolja21 (talk) 22:29, 14 September 2015 (UTC)
- @Kolja21: Ja, das ist möglich. Ich kümmere mich um eine Lösung für beide Probleme. Hast du eine Idee, wie ich den Typ aus der GND auslesen kann? Ich habe etwas in der MARCXML-Datei unter Datafield 079 / Subfield b gefunden. Ist das die richtige Stelle? Viele Grüße, -- T.seppelt (talk) 06:32, 15 September 2015 (UTC)
- Super, danke! Die Satzart wird (bei Pica+) in Feld 002 angezeigt. Bei dem Datensatz für "Burney, Charles Fox" z.B. "Tn6", angelegt von der Bayerischen Staatsbibliothek (DE-12).de:Benutzer:APPER ist der Technikexperte, der sich mit dem Auslesen der Felder auskennt. --Kolja21 (talk) 12:05, 15 September 2015 (UTC)
- Es war tatsächlich 079/b. Ich habe in der Doku nachgelesen. Ich könnte jetzt einfach ein Programm schreiben, dass alle GND werde für instance of (P31)human (Q5) prüft und die, die Typ n sind auflistet. Um das Löschen bei VIAF als Quelle kümmere ich mich. Viele Grüße, -- T.seppelt (talk) 14:21, 15 September 2015 (UTC)
- Das wäre eine riesen Hilfe. Dann können wir hoffentlich auch wieder die Liste Constraint violations/P227 einfacher abarbeiten. (Zzt. werden die labels ausgeblendet, da die Zahl von 500 Fehlermeldungen überschritten ist.) --Kolja21 (talk) 14:42, 15 September 2015 (UTC)
- @Kolja21: Ich habe das Programm gestartet. Es löscht wenn imported from Wikimedia project (P143)Virtual International Authority File (Q54919) als Quelle angegeben ist, siehe dazu Special:Contributions/KasparBot. Den Fehlerbericht gibt es in den nächsten Tagen. Viele Grüße, -- T.seppelt (talk) 14:44, 15 September 2015 (UTC)
- Das wäre eine riesen Hilfe. Dann können wir hoffentlich auch wieder die Liste Constraint violations/P227 einfacher abarbeiten. (Zzt. werden die labels ausgeblendet, da die Zahl von 500 Fehlermeldungen überschritten ist.) --Kolja21 (talk) 14:42, 15 September 2015 (UTC)
- Es war tatsächlich 079/b. Ich habe in der Doku nachgelesen. Ich könnte jetzt einfach ein Programm schreiben, dass alle GND werde für instance of (P31)human (Q5) prüft und die, die Typ n sind auflistet. Um das Löschen bei VIAF als Quelle kümmere ich mich. Viele Grüße, -- T.seppelt (talk) 14:21, 15 September 2015 (UTC)
- Super, danke! Die Satzart wird (bei Pica+) in Feld 002 angezeigt. Bei dem Datensatz für "Burney, Charles Fox" z.B. "Tn6", angelegt von der Bayerischen Staatsbibliothek (DE-12).de:Benutzer:APPER ist der Technikexperte, der sich mit dem Auslesen der Felder auskennt. --Kolja21 (talk) 12:05, 15 September 2015 (UTC)
@Kolja21: das Programm ist zwar abgestürzt, hat aber 1876 Einträge für diese Kriterien gefunden. Die Liste ist unter User:KasparBot/GND Type N. Viele Grüße, -- T.seppelt (talk) 06:19, 19 September 2015 (UTC)
- Höchst plausibel: Im Dump vom Montag waren es 1879 Tn-Sätze. -- Gymel (talk) 06:34, 19 September 2015 (UTC)
- Okay gut. Ich stelle gerade fest, dass das Programm doch noch läuft. Es geht aber die Items absteigend nach ID durch. Bei den niedrigeren IDs, d.h. oft wichtigeren Leuten, ist die Wahrscheinlichkeit für Tns wohl niedriger. Gymel, du scheinst einen effizienteren Weg zu haben, um das herauszufinden. -- T.seppelt (talk) 06:39, 19 September 2015 (UTC)
- Ja, aber nur das. Also nicht in Kombination mit anderen Kriterien wie etwa "importiert aus" oder gleichzeitigem Vorliegen meherer Werte für P227. -- Gymel (talk) 06:49, 19 September 2015 (UTC)
- Okay, vielleicht ist damit Kolja21 schon geholfen. Das Löschen der VIAF-Importe habe ich weitesgehend erledigt. -- T.seppelt (talk) 06:54, 19 September 2015 (UTC)
- Sehr geholfen sogar! In den Fällen, wo der Tn mit einem Titel verknüpft ist, der zu der Person gehört, kann eine "richtige" GND (Tp) beantragt werden. In der deutschsprachigen Wikipedia machen wird das ja schon seit 2005; jetzt können wir auch die anderen Wikis in die Wartungsarbeit mit einbeziehen. --Kolja21 (talk) 11:43, 19 September 2015 (UTC)
- Okay, vielleicht ist damit Kolja21 schon geholfen. Das Löschen der VIAF-Importe habe ich weitesgehend erledigt. -- T.seppelt (talk) 06:54, 19 September 2015 (UTC)
- Ja, aber nur das. Also nicht in Kombination mit anderen Kriterien wie etwa "importiert aus" oder gleichzeitigem Vorliegen meherer Werte für P227. -- Gymel (talk) 06:49, 19 September 2015 (UTC)
- Okay gut. Ich stelle gerade fest, dass das Programm doch noch läuft. Es geht aber die Items absteigend nach ID durch. Bei den niedrigeren IDs, d.h. oft wichtigeren Leuten, ist die Wahrscheinlichkeit für Tns wohl niedriger. Gymel, du scheinst einen effizienteren Weg zu haben, um das herauszufinden. -- T.seppelt (talk) 06:39, 19 September 2015 (UTC)
Fehlerhafter Import
Der Bot hat bei Lerchenköpfe (Q1820245) eine Höhe von 8.210m statt 821m aus der WP importiert. Konnte diesen Fehler nur bei einem Berg erkennen, wäre aber vielleicht interessant zu überprüfen. Ogmios (Tratsch) 05:42, 18 September 2015 (UTC)
- @Ogmios: danke fürs Bescheid sagen. Das Problem liegt irgendwo bei den Java NumberFormats, ist aber einmalig. Im Artikel zum Lerchenkopf war die Höhe mit 821.0 angegeben. Der Doubleparser hat den Punkt wohl nicht als Dezimaltrennzeichen erkannt. Ich arbeite an einer besseren Implementation. Viele Grüße, -- T.seppelt (talk) 14:02, 18 September 2015 (UTC)
- Also einmalig ist das sicherlich nicht. Geschah so bspw. auch bei Keulaer Wald (Q1740023) und Rehhecke (Q2138969). Bitte nochmal über entsprechende Stellen laufen und korrigieren lassen. Grüße --Nattr (talk) 23:03, 19 September 2015 (UTC)
Achtung, wahrscheinlich viele Werte betroffen [17] ist nicht 586,6 m. Grüße, Conny (talk) 20:47, 22 September 2015 (UTC).
Hello, your bot adds bishop (Q29182) to occupation (P106) in many articles (e.g. https://www.wikidata.org/w/index.php?title=Q1694676&type=revision&diff=212154059&oldid=205506986). I think it's not correct, because bishop (Q29182) is a position (Q4164871) not a profession (Q28640) so it should be added to position held (P39). Would you like to fix it? Nurni
- @Nurni: thank you. I take this in consideration when I restart this program. My bot didn't perform this task for a while. Warm Regards, -- T.seppelt (talk) 15:35, 1 October 2015 (UTC)
- Difficult: vs. (I thought it's the other way round). Anywa: almost all persons of religious rank "bishop" (you don't lose that on retirement) perform the religious functions of a bishop and for many of them the full-time occupation is administering a diocese, i.e. being "bishop". -- Gymel (talk) 18:23, 1 October 2015 (UTC)
Can you check this edit? The right value is 3402.2, not 34022. --ValterVB (talk) 17:53, 5 October 2015 (UTC)
- another one [18]. Holger1959 (talk) 13:21, 3 November 2015 (UTC)
- @Ogmios, Nattr, Conny, ValterVB, Holger1959: I checked all pages and corrected the values. I am sorry for this bug. The problem in the framework is fixed. Warm regards, -- T.seppelt (talk) 18:21, 4 November 2015 (UTC)
itwiki on Authority control validator
Hi! I've found your fantastic tool! It will be very useful for us if you also add itwiki. If you need some info about our templates or cats, feel free to contact me. Thanks! --AlessioMela (talk) 12:00, 10 October 2015 (UTC)
- @AlessioMela: thank you for this message. According to Wikidata:WikiProject Authority control/Status itwiki doesn't make use of local parameters anymore. You only call the template with
{{Controllo di autorità}}
. Therefore they can't be any differences or issues in general to track with the tool. Warm regards, -- T.seppelt (talk) 17:36, 10 October 2015 (UTC)
Lists and authority IDs
Hello, please exclude Wikimedia list article (Q13406463) items from processing: [19], [20]. — Ivan A. Krestinin (talk) 19:35, 23 November 2015 (UTC)
- @Ivan A. Krestinin: Until now I decided to not implement this check because removing the information on a Wikipedia and not adding it on Wikidata would look like magic disappearance. I would recommend to remove those statements manually and check it with the Wikipedia where the statement was imported from. -- T.seppelt (talk) 16:13, 24 November 2015 (UTC)
- But in the first example your bot took data from two sitelinks to introduce the erroneous statement three times! -- Gymel (talk) 16:17, 24 November 2015 (UTC)
- Sorry. The log of list of works by Amir Hamzah (Q13567614) looks indeed very strange. It should not happen anymore because the Authority control templates of all three connected Wikipedia articles are blank now, but this is not a general solution. What would you prefer? How should the bot decide? In case of instance of (P31)Wikimedia list article (Q13406463) leave the value at the Wikipedia? -- T.seppelt (talk) 16:22, 24 November 2015 (UTC)
- I wouldn't base any decision on the P31 value because this might be wrong as well (some other sitelink erroneously added to the item, later removed without cleanup is a common situation). Could you detect that you are about adding a specific authority control value for a second time because it has already been removed or changed, and then just put some notice on the item's talk page instead? When most of this is due to KrBot in the meantime having substituted a VIAF value by its redirection target, then of course this would be a very bad strategy. Personally, I'm trying to explicitly undo bot actions even if its more cumbersume than simply deleting a wrong value, if other editors would do the same perhaps that would then be a means to detect unwanted entries? -- Gymel (talk) 17:21, 24 November 2015 (UTC)
- Sorry. The log of list of works by Amir Hamzah (Q13567614) looks indeed very strange. It should not happen anymore because the Authority control templates of all three connected Wikipedia articles are blank now, but this is not a general solution. What would you prefer? How should the bot decide? In case of instance of (P31)Wikimedia list article (Q13406463) leave the value at the Wikipedia? -- T.seppelt (talk) 16:22, 24 November 2015 (UTC)
- But in the first example your bot took data from two sitelinks to introduce the erroneous statement three times! -- Gymel (talk) 16:17, 24 November 2015 (UTC)
- Deleting local values from Wikipedia is error for list articles. It must be replaced to {{Authority control|datasource=Q3981632}} or stay unchanged. — Ivan A. Krestinin (talk) 21:41, 24 November 2015 (UTC)
- Most Wikipedias don't support the arbitrary access feature for Authority control. To use the datasource-parameter is therefore in most cases no option. Putting a notice on the talk page is the better option, I think. -- T.seppelt (talk) 11:46, 26 November 2015 (UTC)
Wrong edition
Just for you information, a bot wrong edition.[21]--Pere prlpz (talk) 23:27, 2 December 2015 (UTC)
- Thank you. T.seppelt (talk) 08:28, 3 December 2015 (UTC)
ULAN import from VIAF
Hi T.seppelt, I thought you had a bot to do imports based on viaf. Am I correct? I would like a bot to go over all items that have occupation (P106)->painter (Q1028181) and VIAF ID (P214), but no Union List of Artist Names ID (P245) like this edit. This would probably help to reduce User:Multichill/Paintings creator no authority control in the future (now it still includes viaf). Multichill (talk) 21:41, 30 November 2015 (UTC)
- Nevermind. I figured it out myself and imported quite a few new ULAN links. Multichill (talk) 21:13, 5 December 2015 (UTC)
- Sorry, I was on vacation. I'm glad to see that the problem is solved. Warm regards, -- T.seppelt (talk) 10:04, 7 December 2015 (UTC)
Enwiki persondata import
Some comments:
- Make the skip button bigger, red, and place it slightly more prominently. (Apparently you had this planned but I'll echo it; "I should make it maybe bigger" from November 24.)
- Make unavailable the varied data which indicate a date earlier than 1920, per the brief discussion at Wikidata talk:Primary sources tool#Migration of enwiki Persondata. Or move it into a different workflow, or something. Earlier dates are a mess right now and I don't think this tool should exacerbate that fact.
- You can probably improve the link to the article by using the Wikidata item's link rather than using Special:Search+Article+Name.
- It might be nice to have a link available to the version of the article at the date of import. This way I can take a look to see if anything in the article is particularly disagreeable to the persondata as well as how different (whether for example a date is a refinement of the date elsewhere in the article, see e.g. [22] or more likely to have vandalized). On this point, a link to the history of the article would also be appreciated.
Thanks for the tool. On the primary sources talk page, it looks like the entire dataset was removed from the primary sources workflow. Is that correct? --Izno (talk) 15:26, 11 December 2015 (UTC)
Also, there are a number of challenges where the page on the Persondata side is a disambiguation item and the other is an actual item where the titles on Wikidata are the exact same. Maybe these can be prefiltered? --Izno (talk) 16:24, 11 December 2015 (UTC)
- Moin Moin T.seppelt, ist es vorstellbar, dass du dieses auch für die deutsche Personendaten-Vorlage machst? mfg --Crazy1880 (talk) 12:38, 3 January 2016 (UTC)
- Moin, ja, vorstellbar ist das schon. In der Deutschen Wikipedia sehe ich aber keine Bewegung was das Einstellen des Persondendatenprojektes anbelangt. Das ist es, worum es bei dem Tool geht. Die Verwertung der Reste aus den Persondata-Beständen während die Daten aus den Artikeln gelöscht werden. - - T.seppelt (talk) 13:06, 3 January 2016 (UTC)
- Moin Moin T.seppelt, ist es vorstellbar, dass du dieses auch für die deutsche Personendaten-Vorlage machst? mfg --Crazy1880 (talk) 12:38, 3 January 2016 (UTC)
KasparBot - thank you
Just a note to say: thank you for the hard work you've put in on KasparBot. I've just started going through the Persondata imports and trying to help with the descriptions. There's so much to do, and it is only through an intelligent combination of human curation and software that we will make Wikidata useful. Great work. —Tom Morris (talk) 19:00, 20 December 2015 (UTC)
- +1 ;) --Kolja21 (talk) 00:57, 21 December 2015 (UTC)
Extremely slow interface for challenges
Right now the challenges pages only show one item at a time, and it takes 4–6 seconds to load the next page each time I click one of the decision buttons. Perhaps this could be solved by using AJAX and showing multiple items per page, so I can make a decision and immediately see the next challenge. In any case, the way it is now makes it difficult to work very effectively on the queue. StephenWade (talk) 23:50, 1 February 2016 (UTC)
- I actually found that you can open two tabs and effectively cut the wait time to zero by flipping between tabs. That's probably a workaround to the core issue, but you should give it a try. --Izno (talk) 12:49, 2 February 2016 (UTC)
- @StephenWade, Izno: I am using the same workaround. I did some tests with the loading time. The main issue seems to be the database which is extremely slow. Showing several challenges at once would probably reduce the loading time but would also make the interface more complicated. The users should probably be able to decide how many challenges they want to see at once... Another issue is the time which is needed to perform the edit after making the decision. The Wikidata servers take some time to finish the whole process. I don't really want to come up with a system which communicates with the Wikidata servers in the background. This would make the server side script much more complicated and error-prone. -- T.seppelt (talk) 15:06, 2 February 2016 (UTC)
Basic question
Over the past few days, I've seen items on my watchlist being edited by KasparBot. There also seems to be some type of verification "challenge" associated with them. I do like a good challenge, but I simply don't understand what is being done here and what is involved with "authorizing" KasparBot. Might you point me to some page or article that will explain this? Thanks. NewYorkActuary (talk) 22:37, 3 February 2016 (UTC)
- @NewYorkActuary: Thank you for your question. First a couple of basic things: The current task of KasparBot is to delete all Persondata usages. The Persondata template was deprecated by an RfC. It provided structured information (name, description, place and date of birth and death) about people in the English Wikipedia. Wikidata takes care of these information know (descriptions, aliases, date of birth (P569), date of death (P570), place of birth (P19) and place of death (P20)). Because a part of the data is still usable it was copied to a database. Users can use this page to transfer the mentioned information manually to Wikidata. Everybody is kindly ask to help. As first step you need to authorize the page. You can find information about this on mw:Help:OAuth. I'd recommend you to follow the links. Click on Authorize, than on the green accept button and than on places for instance. If you have more questions, just drop me a note. Warm regards, -- T.seppelt (talk) 16:09, 4 February 2016 (UTC)
Questions about choosing the right correction
I recently came across challenges to Donald Smith, 1st Baron Strathcona and Mount Royal (https://www.wikidata.org/wiki/Q337004) and determined from an authoritative source that Wikidata's date of death, 1914, was right and Persondata's 1915 was wrong. But I must have found the interface confusing, because Wikidata recorded the opposite of what I intended; see the history page's current 4th and 5th entries. Since WD still said 1915 I wondered if there was a delay or a step I had missed, so the first question is: are the changes made to WD immediately?
I waited a few minutes and re-corrected the data (and place) by hand. Of course, now I've dealt with the claim (claim:28712) I can't reproduce the interface, but I thought I had done the right thing. I'm trying to find a similar "date-of" claim to test my assumptions about the UI, and so that I can suggest a clearer instruction wording. I'm also wondering if the UI was impacted by the shape of the browser window.
One suggestion: it would be helpful if the Browse Challenges page (e.g. https://tools.wmflabs.org/kasparbot/persondata/challenge.php?q=article%3ADonald+Smith%2C+1st+Baron+Strathcona+and+Mount+Royal) could contain the name of the property alongside its P number, as it does on an individual claim challenge page like https://tools.wmflabs.org/kasparbot/persondata/challenge.php?q=claim%3A28712.
I'm not a dedicated claim-challenger; I was aware of the Persondata migration, and this one just happened to be a recent change in an article I was editing for other reasons. DavidBrooks (talk) 22:27, 23 February 2016 (UTC)
- @DavidBrooks: As far as I can see you did completely the right thing. The changes which originate from the decisions are applied immediately. This was also the case at Donald Smith, 1st Baron Strathcona and Mount Royal (Q337004). I fixed the bug with the missing labels at challenge.php. Now the names of the properties are also displayed. Thank you for helping with the migration, -- T.seppelt (talk) 06:06, 24 February 2016 (UTC)
- @T.seppelt: Actually I think I did the wrong thing at first (this was the change resulting from using the challenge UI) but later fixed it by hand. I'll let it go until I come across another example, and will try to see what seemed to me to be ambiguous here. DavidBrooks (talk) 23:11, 26 February 2016 (UTC)
- @T.seppelt: Now I can see the problem. I didn't understand the phrase "No, this is enough!". It means "Wikidata is complete and correct"; how about using a phrase like that? Also, I think the first button would more logically be something like "Remove the current value and use this". Finally, "Add this" would be an unusual choice (two equally valid dates of death?) so perhaps it should not be green. DavidBrooks (talk) 06:10, 2 March 2016 (UTC) (ETA) I realize English isn't your first language, so I hope this is helpful. DavidBrooks (talk) 06:17, 2 March 2016 (UTC)
- This is a good point. I'll adept the UI. You are right, English is not my native languages. Thank you for the help. --T.seppelt (talk) 15:38, 2 March 2016 (UTC)
- Bitte! DavidBrooks (talk) 16:44, 2 March 2016 (UTC)
- @DavidBrooks: I changed the UI based on your suggestions including labels and colours. Warm regards, -- T.seppelt (talk) 19:57, 7 March 2016 (UTC)