User talk:Pxxlhxslxn

Metropolitan Museum of Art

Why did you create Metropolitan Museum of Art (Q121079403)? Seems to be a duplicate of Metropolitan Museum of Art (Q160236). Multichill (talk) 16:31, 1 August 2023 (UTC)

Thanks for picking up on this, this is now fixed. I had MMA as a production company which I had thought could have been distinct from the art institution, but these entities are now merged. Pxxlhxslxn (talk) 19:53, 1 August 2023 (UTC)

I'm thinking Australian Museum (Q121058106) and Australian Museum (Q769416) should also be merged. Azertus (talk) 16:16, 22 August 2023 (UTC)

Thank you for noticing this @Azertus, these have now been merged. Pxxlhxslxn (talk) 21:58, 22 August 2023 (UTC)

Please stop editing

You have imported a ton of low quality items and are now incorrectly merging items. Please stop and explain what you are doing or you will be blocked. BrokenSegue (talk) 17:22, 11 August 2023 (UTC)

Good morning @BrokenSegue, happy to provide context to the work being undertaken. I am currently working as the Wikimedian in residence working with ACMI at integrating their extensive public catalogue of Australian audiovisual creators and artistic works into Wikidata, which involves adding entities where they do not previously exist. I realise these entities currently exist as data “islands” as we have not added filmographies yet, but this would be the next phase of the project. The two items you have raised as incorrect merges, I don’t believe are incorrect as such - the entities are truly the same, the issue is in inferring the entity type from source data, where in this case the “composer” database statement is almost always applied to a human, but there are exceptions which you have identified and I will get to repairing. Pxxlhxslxn (talk) 21:47, 11 August 2023 (UTC)

It seems that being a Wikimedian-in-Residence is a notable and relevant fact that could be announced on your user page, along with your current project, as you have stated it. Elizium23 (talk) 22:25, 11 August 2023 (UTC)

Thank you for the quick response, I will update this later today with some more extensive information around the scope of the project and methods undertaken. Pxxlhxslxn (talk) 22:30, 11 August 2023 (UTC)

In this edit you labeled London Symphony Orchestra (Q155805) as a human (Q5). That is objectively wrong. Also, your imported items all lack descriptions or other basic information. I'm glad to hear you plan to go back and fix this up though.

Are you using a bot to do these merges? If so it does not appear to be respecting maxlag given its edit speed. BrokenSegue (talk) 23:04, 11 August 2023 (UTC)

sorry I meant are you using a bot to do these item creations. BrokenSegue (talk) 23:04, 11 August 2023 (UTC)

@BrokenSegue I completely agree, as stated above this is due to trying to infer “humanness” based on source data contribution to artistic work, and clearly this is not holding true in all cases. I am not using a bot as such (by which I would consider a completely autonomous unit), I am feeding tables of data of around 1000 rows. I apologise for not configuring maxlag correctly, I will make that change. Pxxlhxslxn (talk) 23:12, 11 August 2023 (UTC)

I appreciate your work but re this would be the next phase of the project: I see two problems with that:

1. Wikidata doesn’t work like that. Items should “work” at all stages of their history.
2. I have seen many similar projects, some of very experienced people from prestigious institutions, that have never reached a promised “second phase” – because of technical problems, lack of funding, unforeseeable changes in their proponents’ life plans or simply because of inertia. Your project might be different and I sure hope so – but please understand that I’m wary when somebody creates 11,672 items that may or may not be fixed in the future. That’s a lot of work for other users if for some reason it doesn’t work out.

--Emu (talk) 23:12, 11 August 2023 (UTC)

@Emu With this in mind I will look at shifting focus on bringing these entities up to standard asap before creating any new ones. Pxxlhxslxn (talk) 23:14, 11 August 2023 (UTC)

Thank you! --Emu (talk) 23:51, 11 August 2023 (UTC)

Kirk Smalmann

Hello Pxxlhxslxn, there is also a Kirk Smallman (Q22101682). Is that the same? In that case they should be merged, I guess. Cheers. Lotje (talk) 06:25, 13 August 2023 (UTC)

Thanks for picking this up @Lotje, and it does seem likely that these producers are the same person, but I cannot find a resource which explicitly links them together. Pxxlhxslxn (talk) 04:26, 15 August 2023 (UTC)

No worries. Thanks anyway :-) Lotje (talk) 10:02, 15 August 2023 (UTC)

Both were in the production company Henry Street Settlement in the same decade: [1] and [2]. I'll merge. --99of9 (talk) 05:08, 28 August 2023 (UTC)

P7003 takeover

Hi Pxxlhxslxn,

As the original proposer of the property ACMI ID (P7003), I'm going to take your unilateral redefinition and takeover of the property as a friendly amendment (Q5504090). However, I'd like to gently suggest that this should not be the way things are done around here. Such an important redefinition should have been discussed first at the property talk page. The initial structure had already been agreed by over a dozen contributors, so even if you represent the organization depicted by this property, collaboration and explanation is expected. For example, some questions I still have:

• why "work" and "creator" IDs should not be separate IDs, upon which we could easily put separate and more targeted constraints.
• the current status of the collection "object" ID (which the property originally used):
  • Is there a reason why it is no longer valid and is now used as a web redirect?
  • Is it expected to be maintained as a persistent ID? Have you mapped across all community matches of this original ID?
  • Have you considered what will happen when contributors continue putting in "object" IDs from the existing Mix'n'Match catalogue? Perhaps we should have discussed deprecating that catalogue?
• How do you expect further matches to be made after your project finishes?
  • Are you planning to set up a title and creator Mix'n'Match catalogue so the community can monitor the progress and make manual incremental matches?
  • What will be the ongoing role of your scripts or your bot? Please can you detail the tasks it will take on in a Bot request?

I'm glad to see bits of the project explained at WMAU and github, and look forward to more. I very much appreciate your efforts on matching and uploading new items about these people and works, as well as the way the ACMI is linking back to and reusing the Wikidata metadata. I think this has great potential going forward, as long as the surge of project work fits well into an ongoing plan.

Best regards, --99of9 (talk) 06:30, 28 August 2023 (UTC)

Good evening @99of9,

I had not realised you originally initiated this property, so good to remeet you within this context! Thank you for pointing out that major property rescoping should be discussed prior to any action, I will certainly follow this advice for any similar modifications in the future.

To address your specific concerns:

-- Individual properties: I am absolutely happy to follow your lead on transferring to distinct properties, if you feel like this is the preferred pattern.

-- Legacy URLs: The collection object link pattern was broken from a website redesign a short period back before I started, and in fact when I came to this project there were no redirects in place, so all of the P7003s linked to an ACMI 404.

-- Persistence: It is expected that the current pattern will persist, or at the very least there is increased awareness that redirects are required to maintain legacy connections. All legacy links were migrated to the new pattern.

-- Mix’n’Match: I have to confess I have not used Mix’n’ Match myself, but it looks like an excellent option for the many outstanding works which I was hoping to get through within the next few weeks. How would I go about providing a fresh API export to populate the catalogue? And would you recommend tweaking the existing instance or starting fresh?

-- Ongoing matches: This is a current subject of some discussion, I had published an OpenRefine repo which identifies unmatched entities and formats for that tool, but I am now wondering if MnM would be a better platform as it requires no extra installation? There is a bot in effect already, although this works only by writing internal matches which are created by staff outwards, not the other way around. The best method for generating new works with attached data statements would be something I would be keen to discuss further, if you could suggest the best forum for such a conversation.

Thank you also for your interest and words of support, I am just finishing a blog post which should hopefully provide much more detail on the different approaches taken and lessons learned from this residency.

all the best, Pxxlhxslxn (talk) 09:20, 28 August 2023 (UTC)

Thank. Openrefine is great too. They work differently and suit different contributors, so it would be great to have both methods working. All that's needed for a MnM set is a csv/tsv with columns for the ID, the name, and a description. I'd be happy to help with the upload setup if it helps. Alternately, to make a dynamic set, a web scraper can be set up. This is doable but significantly more complex. Either way we should certainly start fresh since the old IDs are deprecated (so I'll also ask for the old set to be deactivated). --99of9 (talk) 01:04, 30 August 2023 (UTC)

@99of9 Sorry for the length delay, but I have now put together a first pass at a Mix'n'Match dataset here. Welcome for any feedback, especially around the description formatting. I have reduced to Films for now, as this is still over 35K rows. If you are ready to disable the previous instance, I can set this now one up. All the best, Pxxlhxslxn (talk) 06:08, 23 September 2023 (UTC)

Good job. My only suggestion is to be careful with UTF-8 encoding. For example line 35863 shows up to me as "118342 Grace Tame’s address to the National Press Club 2021 film". I'll ask for deactivation of the other set now. --99of9 (talk) 00:54, 25 September 2023 (UTC)

I have added the new catalog as 6044, but an issue which is not immediately obvious for me how to resolve - how do I delete previous imports and start fresh? My first pass had incorrectly formatted work ids, which I have corrected in the git repo and would like to reimport as the catalog. Pxxlhxslxn (talk) 19:43, 26 September 2023 (UTC)

Recent movies import

Hi Pxx,

are you gonna import all IMDB movies? How have e.g. chosen Cuckoo Waltz (Q122661846) as eligible?

Kind regards-- U. M. Owen (talk) 19:51, 19 September 2023 (UTC)

Good evening, no, I am specifically adding films cataloged in Amos Vogel's Film As A Subversive Art book, and have just added IMDB ids where they already existed. Best, Pxxlhxslxn (talk) 20:00, 19 September 2023 (UTC)

Bionicle: Mask of Light

Q123373622 is a duplicate of Q975962. --Jaller94 (talk) 05:35, 23 November 2023 (UTC)

This has been merged, thank you for spotting it. Pxxlhxslxn (talk) 09:26, 23 November 2023 (UTC)

NFSA title ID

Hi there! Just a heads up, there's a difference between identifiers for individual tracks (in this case the title track) and the albums that contain them, see Howzat. Moebeus (talk) 17:55, 29 October 2024 (UTC)