Wikidata:Requests for permissions/Bot/DennyBot
- The following discussion is closed. Please do not modify it. Subsequent comments should be made in a new section. A summary of the conclusions reached follows.
- Withdrawn by the bot operator. --Ricordisamoa 03:15, 7 December 2013 (UTC)[reply]
DennyBot (talk • contribs • new items • new lexemes • SUL • Block log • User rights log • User rights • xtools)
Operator: Denny (talk • contribs • logs)
Task/s: uploading data for Property:P646, i.e. Freebase identifiers, based on the CC0 dataset released. I withdraw my request to implement this task.
Function details: see Tasks. It is pretty simple. I can make a few test edits with the bot, if preferred. I am using PyWikipedia as a Bot-Framework. The DennyBot account has been used for a long while by me for other tasks on a small number of Wikipedias. By putting 'stop' on User:DennyBot/Emergency, anyone can stop the bot. --Denny (talk) 22:49, 18 November 2013 (UTC)[reply]
- Could you run a trial? The tasks seems pretty much non controversial apart from users objecting to us storing Freebase identifiers. Test edits of about 50? John F. Lewis (talk) 22:54, 18 November 2013 (UTC)[reply]
- Sure. --Denny (talk) 22:56, 18 November 2013 (UTC)[reply]
- Hi Denny, the interesting part is: how will your bot source the addition of Freebase ID (P646) :) --Succu (talk) 22:57, 18 November 2013 (UTC)[reply]
- Indeed! Hmm. It could point to this URL [1], which describes the mapping file and provides it. Would that be an acceptable solution? --Denny (talk) 23:57, 18 November 2013 (UTC)[reply]
- I think linking to an URL is enough, but it should have a stable content (that is, if you make a new version of the mapping, it should still link to the version that was used by the bot). --Zolo (talk) 08:39, 19 November 2013 (UTC)[reply]
- Indeed! Hmm. It could point to this URL [1], which describes the mapping file and provides it. Would that be an acceptable solution? --Denny (talk) 23:57, 18 November 2013 (UTC)[reply]
- There is a paragraph Citing. So I think the best solution is:
- Create a new item and name it Google, Freebase Data Dumps
- Add reference URL (P854) with https://developers.google.com/freebase/data to this item
- Then you can use this item in your reference:
- stated in (P248) = Google, Freebase Data Dumps
- retrieved (P813) = date
- --Succu (talk) 09:53, 19 November 2013 (UTC)[reply]
- There is a paragraph Citing. So I think the best solution is:
- Looking on Wikidata's page on sources, we should not create an item for the Webpage, but rather use reference URL (P854) directly, and then original language of film or TV show (P364) and P357 (P357), publication date (P577) and publisher (P123). I can do that, it looks a little bit like overkill, but I am fine with it. Any thoughts? --Denny (talk) 00:16, 20 November 2013 (UTC)[reply]
- Re-thinking about it, my personal opinion :
- it is always a good idea to directly cite the url and the date retrieved when citing a web page
- the rationale for using the page title rather than stated in (P248) is that it would be overkill to create an item for every cited webpage, but it would make sense to have an item for a source cited several millions times.
- language does not seem really useful when citing a database with not much natural language in it.
- So actually, I would agree with Succu except that I would prefer to add the URL to directly to the source (perhaps the URL for the source can be made more accurate than the main URL of the item). So something like stated in (P248): Google Freebase dump + retrieved (P813) (date when the data was extracted) + reference URL (P854). --Zolo (talk) 08:18, 20 November 2013 (UTC)[reply]
- Re-thinking about it, my personal opinion :
- OK, check out quark (Q6718) where I did the source by hand. I also created the item for the Freebase dumps, Freebase Data Dumps (Q15241312). If no one complains, I will go ahead to make the bot do this kind of sources to the IDs. --Denny (talk) 18:28, 27 November 2013 (UTC)[reply]
- I think it's redundant to repeat reference URL (P854) and publisher (P123) at every reference. --Succu (talk) 18:52, 27 November 2013 (UTC)[reply]
- OK, I will remove publisher (P123). reference URL (P854) though does point to a more accurate URL than the item, as per Zolo's suggestion above. Shall I still remove it? --Denny (talk) 20:05, 27 November 2013 (UTC)[reply]
- Denny, in my opinion you should simply rename Freebase Data Dumps (Q15241312) to something like Freebase Wikidata Mappings and change URL to https://developers.google.com/freebase/data#freebase-wikidata-mappings. Then all is fine. :) --Succu (talk) 20:12, 27 November 2013 (UTC)[reply]
- Easy to do. If no one disagrees until Monday, I will make the changes suggested by Succu and then the 50 test edits. --Denny (talk) 20:41, 27 November 2013 (UTC)[reply]
Oppose As said above "non controversial apart from users objecting to us storing Freebase identifiers." There is no consensus for this task. Frankly it's awful how google is pushing this into wikidata. See relevant discussions at:
- Wikidata:Property_proposal/Archive/8#Freebase_identifier
- wikidata-l june 2013
- wikidata-l nov 2013
- Wikidata:Project_chat/Archive/2013/11#Freebase-Wikidata_mappings
- Wikidata:Bot_requests#Wikidata_Freebase_mappings
- Property_talk:P646
- de:Wikipedia:Projektdiskussion/Wikidata#Wikidata und Google. --Atlasowa (talk) 23:09, 30 November 2013 (UTC)[reply]
- Hello Atlasowa. Thank you for your comments. It is absolutely correct that the Freebase identifier property and the mapping to Freebase were discussed in several places. It is my understanding that this is the appropriate way for getting things done here. In the future, I intend to keep discussing such proposals before implementing them. As you can see in the previous discussions, this has lead to refined and improved proposals. Thank you for going through the work of collecting the references. Let us look at them:
- A number of comments, many of them positive. All votes were cast in support.
- Two questions to Shawn on the mailing list. That thread had more than a dozen replies. The posting you selected does not even discuss the relevant property.
- This is the mailing list discussion after we released the mappings.
- This is the on-wiki discussion after the release of the mappings.
- This is the actual place where this task should be discussed, not here. This is the discussion about the technicalities of my bot and how it should run in order to implement the task. I would suggest to move your argument to this place. There are already five supports, and one wait regarding this task.
- This is the talk page of the property. There is one person asking about the usefulness of this property (which was answered), and one pointing to this discussion.
- de:Wikipedia:Projektdiskussion/Wikidata#Wikidata und Google. --Atlasowa (talk) 23:09, 30 November 2013 (UTC)[reply]
- This is the discussion you have started on the German Wikipedia, and where you question the viability and quality of Wikidata as a whole.
- I would suggest to move this discussion to the discussion of the task as such, not here. Here we should discuss the technical details of implementing the task, given that the task is accepted. I am happy to discuss these here, but I don't actually see anything in your argumentation pertaining to that. --Denny (talk) 19:42, 3 December 2013 (UTC)[reply]
Oppose I don't see much use in Freebase and it really would be nice if Google would stop pushing this property. --Kolja21 (talk) 09:12, 1 December 2013 (UTC)[reply]
- Hi Kolja21. As said above, this page is about DennyBot implementing the task of uploading the mappings to Freebase. Can you please explain how your oppose relates to this? If you want to discuss the task itself (whether it should happen or not), I guess Wikidata:Bot requests#Wikidata Freebase mappings would be the more appropriate place. If you want to suggest to delete the property itself, Wikidata:Properties for deletion is the right place. --Denny (talk) 19:42, 3 December 2013 (UTC)[reply]
- Basically, both opposes are opposing this task because they feel Google is pushing Freebase on Wikidata? No. A Google employee did propose the property be created, but they did not force the community to create the property. Google did create mappings for Freebase to Wikidata, but they are not forcing us to use it. Both opposes are literally assuming and faith against Google who on a side note, did fund the development of Wikibase/Wikidata and without said funding meaning Wikidata may not be where it is today. So, assuming bad faith on Google is completely inappropriate here. Wikidata is tasked to be the central knowledge for Wikimedia and as I know the development team want it, to be a central knowledge for the world. We can not and should not go 'Sorry, it is Google' and then oppose everything relating to them otherwise we are failing the goal of Wikidata and most likely limiting all the work the developers do to bring us Wikibase. Right, so can we please get over the fact Freebase is developed by Google and simply get on with being a centralised knowledgebase for the world? John F. Lewis (talk) 18:57, 3 December 2013 (UTC)[reply]
- @John: As you said: "A Google employee did propose the property be created" and an other Google employee ask to run a bot to make it more popular. The answer is a clear no! That has nothing to do with assuming bad faith. It's the same as in Wikipedia: Can I write about my company? I only want to improve Wikipedia! The answer: en:Wikipedia:Conflict of interest. As long as there are users that think that they have to be thankful for a company sponsoring a project that helps to improve their search engines, the independence of Wikidata is at risk. --Kolja21 (talk) 19:50, 3 December 2013 (UTC)[reply]
- @Kolja21: is this specifically about me, being a Google employee, running the bot? --Denny (talk) 21:04, 3 December 2013 (UTC)[reply]
- @John: As you said: "A Google employee did propose the property be created" and an other Google employee ask to run a bot to make it more popular. The answer is a clear no! That has nothing to do with assuming bad faith. It's the same as in Wikipedia: Can I write about my company? I only want to improve Wikipedia! The answer: en:Wikipedia:Conflict of interest. As long as there are users that think that they have to be thankful for a company sponsoring a project that helps to improve their search engines, the independence of Wikidata is at risk. --Kolja21 (talk) 19:50, 3 December 2013 (UTC)[reply]
One way or the other, if there are any open questions or arguments, I am looking forward to discuss or answer them. I will continue to make proposals and follow the community's decisions, and as far as I can tell, the community has decided that this task should happen. I will obviously wait a day or two before doing the test uploads, in order to provide everyone with enough time to express arguments or questions. --Denny (talk) 19:42, 3 December 2013 (UTC)[reply]
- Hi Denny. Provídes Freebase a backlink to Wikidata? I couldn't find one. --Succu (talk) 21:34, 3 December 2013 (UTC)[reply]
- Not yet, working on it to upload them. --Denny (talk) 21:40, 3 December 2013 (UTC)[reply]
Oppose I subscribe to Kolja21's point of view. It is quite annoying indeed to find a Google employee start a discussion like this in the first place. No use for Wikidata nor for any other Wikimedia projects.--Aschmidt (talk) 22:03, 3 December 2013 (UTC)[reply]
- Thank you for being frank, Aschmidt. I sure do not want to be annoying, and I would appreciate if you would let me know how, in your opinion, a code of conduct for company employees on Wikidata should look like. --Denny (talk) 22:21, 3 December 2013 (UTC)[reply]
I pull back my request to run a bot to implement this task, because it is obviously controversial and to avoid drama. I would still like to suggest that the community comes up with a guideline on what is appropriate for an employee of a company to do. I do not think that the cited en:Wikipedia:Conflict of interest is the actual problem - but maybe I am wrong, and in this case I would appreciate if someone would explain me why I am wrong, and why it indeed constitutes a problem according to enwp's CoI policy. Wikidata is different than Wikipedia, and I think deserves its own policy for that topic. --Denny (talk) 23:16, 3 December 2013 (UTC)[reply]