Samwilson

Tom Brown's School Days

You have broken all the wikilinks between Wikisource and all other related projects on this book. Wikisource has only one edition available, and no other page available for linking. --EncycloPetey (talk) 06:29, 24 July 2015 (UTC)

Sorry if I've annoyed you. ("Unless there is clear evidence to the contrary, assume that people who work on the project are trying to help it, not hurt it." I'm not trying to do the wrong thing here!) Anyway: the link to Wikisource was not broken, but moved to Q20726672. That item is for the 1911 edition that's on Wikisource. Linking directly to the work (i.e. Q7815084) only makes sense if there's only one edition; for example, the publication date of the work is 1857, but the publication date for the Wikisource version is 1911. There should be an item for the work, and one for each of the editions. At least, that's how I read Help:Books. If I'm wrong, sorry! Please revert what I've done. Samwilson (talk) 07:43, 24 July 2015 (UTC)

Reverting isn't possible with the creation of a new item. The issue is not how things are supposed to be done on Wikidata, where I am out of my depth, but how best to coordinate between projects. If a target is removed from all other links on Wikidata because it has been given its own item, the affected project ought to be notified, so that they can ensure links from them and to them are maintained.

In this instance, there is no general page on the English Wikisource, which was why I added it to the existing data item for the work in general. If we had a general page, and separate editions, it would all work as you have done, and indded I have followed some of that myself when dealing with translations of ancient Greek plays. But in this instance, there is only the one item on Wikisource, so isolating it as a data item in itself eliminates all possibilities of interlinkage. Wikipedia will never have an article about the 6th edition of the book; Wikiquote will never have a list of quotations peculiar to that edition; and none of the other Wikisources will ever have such an item either, because the 6th edition of a French translation would not be the same as the 6th edition in English. So, I don't see what purpose is served by having the 6th edition of the English original moved to a data item all on its own.

We on Wikisource have tried to collaborate with Wikidata mavins to sort issues like this out, but become lost in a mire of technobabble, and never know whether our concerns were understood, so I am grateful in this matter to encounter someone speaking to me in coherent, understandable English.

To put the question in another way: Should the English translation of the play Alcestis by Euripides be placed on a separate data item from the item for the original Greek text? If so, will there then need to be a separate data item for every Wikisource, since each has the play in a different language, by a different translator, in a different year? Does this mean that Wikisource should never be linked to the Wikipedia articles unless they are in the original language?

And should the English Wikipedia article on Tom Brown's School Days be removed as well, since it is an article not only about the first edition, but about all editions and even the movie adaptations?

I believe the only sensible interpretation is that a general data item for each work must exist on Wikidata, and links must be there by default. The "date" associated with that item is the original year of publication, and does not necessarily indicate that the item refers to the first edition. There must be a separate data item in existence for the first edition. Editions at Wikisources should only be moved to these edition items if that Wikisource has a general page as well, otherwise we lose all interwikis. --EncycloPetey (talk) 22:27, 24 July 2015 (UTC)

Unused property

This is a kind reminder that the following property was created more than six months ago: inHerit Place Number (P2618). As of today, this property is used on less than five items. As the proposer of this property you probably want to change the unfortunate situation by adding a few statements to items. --Pasleim (talk) 19:27, 17 January 2017 (UTC)

@Pasleim: Thanks for the reminder! I've now got a CSV that I can use. Will try to get some done. Sam Wilson 23:46, 17 January 2017 (UTC)

Cape Cod (Thoreau)

So, even the original edition isn't linked from the primary data item now? --EncycloPetey (talk) 05:19, 9 March 2017 (UTC)

To clarify: Do you mean that the first (original) edition is no longer included in the primary data item for a work, but is given a separate data item? If so, then this needs to be clarified for the entire Wikisource community, since it will have far-reaching ramifications, and will affect hundreds of data items. --EncycloPetey (talk) 05:29, 9 March 2017 (UTC)

@EncycloPetey: (Sorry, I was still updating things. Thanks for adding the edition number.) Yes, the primary (work) item shouldn't have any edition-specific data such as links to scans etc. It should only contain data that is common to every edition. I'm looking for clarity about the use of inception (P571) to indicate date of first publication. And I'm not sure I'm doing anything unusual here am I? I thought this was how it was meant to be done: Wikisource deals almost exclusively in editions, and so it shouldn't impact anything we do there about displaying data etc. and even when we want work-level data (such as sitelinks to Wikipedia) then that's simple enough to get via edition or translation of (P629). It seems reasonably straight forward, and I think is how people have been doing things. (Basically, it's only Wikisource disambiguation pages that should link to work-items here.) Sam Wilson 05:35, 9 March 2017 (UTC)

This isn't the way it's been done up until now for first editions at the English or French Wikisources (at the least), and that's why clarification is needed. One result of this is that there will be no interwiki links between articles about works at Wikipedia and copies of those works on Wikisource, without the addition of special templates at both sites: Adding interwikis directly to the pages doesn't work anymore because well-meaning but misguided editors / bots now routinely strip all interwikis at both projects, assuming that the links are set up with the corresponding data items at Wikidata.

Getting sitelinks using edition or translation of (P629) is simple only for experienced users of WP/WS/WD. For the majority of users, it is neither simple nor at all obvious. Even for experienced users, it means an exploratory scavenger hunt through several pages (with lots of scrolling) to even determine whether or not there will be an article / edition at the other end. --EncycloPetey (talk) 05:43, 9 March 2017 (UTC)

On a related note, how do we mark an edition as "first English edition (original)" versus "first French edition (translation)"? This won't be obvious at all from the primary data item and will almost certainly confuse some users. --EncycloPetey (talk) 05:49, 9 March 2017 (UTC)

@EncycloPetey: Hm, yes, I see what you mean. :-( Okay, so do you agree that it makes sense to have separate items for each edition, and one for the work? Ignoring all usage considerations for a moment, I mean. That much at least seems clear from WikiProject Books. So we need to fix up the systems of linking on other sites? Links from Wikisource templates on English Wikipedia all still use page names, so far as I can see; I'm not sure about other Wikipedias. There are a few templates on English Wikisource (notably {{plain sister}} and similar) that can reasonably easily be fixed to traverse through edition or translation of (P629) — it sounds like that's the next step in this process, do you agree? Or am I getting it all wrong, and should just leave things as they are? My motivation is that I want to be able get associations between e.g. Internet Archive scans and Commons files, and that can't be done unless there is an item for every edition (some of which will point to multiple IA scans and multiple Commons files). Again, I'm really sorry if I'm muddying things! Sam Wilson 06:16, 9 March 2017 (UTC)

This isn't an issue I expected to arise, so I will need to give the question some thought. I do think it would be silly to have separate data items where there is only a single edition extant in the original language, or where the "work" is itself an edition (such as the First Folio of Shakespeare).

We really need a common linking solution to be in place across all WP/WS projects. If we do not have this, then we will have even more frustration and anger at the Wikisources than we currently have (and if you check my talk page, you'll see I get periodic complaints about the current state of affairs even without worrying about editions).

I'm not sure that we will be able to get a clean set of connections between IA scans, Commons files, and Wikisource editions. There are instances where scanned copies include more than one volume or work, where the volumes were published in different years, or where Wikisource amalgamates separate scans into a single work. In short, the linkages are not one-to-one, nor are they likely to be. It is this Wikidata insistence on one-to-one linking that leads to most of the issues I've come across. --EncycloPetey (talk) 07:20, 9 March 2017 (UTC)

original language of film or TV show (P364) is dead

meant to use language of work or name (P407). Went through a deletion discussion and is deprecated for written works.  — billinghurst sDrewth 01:29, 29 August 2017 (UTC)

@billinghurst: Ah, yes I think I remember that now! Thanks for fixing it. :) Sam Wilson 01:34, 29 August 2017 (UTC)

w:Template talk:Marriage

Can I get you to comment one way or the other at w:Template talk:Marriage, I do not care which way you vote, but the issue affects how Wikidata imports data from infoboxes for new entries. --Richard Arthur Norton (1958- ) (talk) 21:06, 27 October 2017 (UTC)

New page for catalogues

Hi, I created a new page where I started collecting sites that could be added to Mix'n'match and I plan to expand it with the ones that already have scrapers by category. Feel free to use, expand. Best, --Adam Harangozó (talk) 10:04, 17 October 2019 (UTC)

State v. State

You might appreciate this query. --99of9 (talk) 05:45, 21 November 2019 (UTC)

Ha! That's terrific! :-) Sam Wilson 05:55, 21 November 2019 (UTC)

Fremantle coordinates

Yes, I do know why they are appearing in lines! If you look at the longitude coordinates, the lines correspond to the third decimal place (115.743, 115.744, 115.745, etc.) or 0.001 degrees, so this is an artefact of the coordinates being rounded to three decimal places. This is really only apparent in an area of high point density such as a major city CBD like Fremantle or Perth. The way I derived the coordinates was to calculate a weighted centroid of the polygons in the Heritage Council WA Shapefile, so I could certainly increase the rounding level to 4 or 5 places and re-import to Wikidata which would mostly address this. --Canley (talk) 02:00, 29 November 2019 (UTC)

I just had a look at the import and for some reason the latitude is rounded to 4 decimal places but the longitude is only 3, which is why the lines are appearing more obviously vertically. I have the data to 8 decimal places so it should be pretty easy to increase the rounding magnitude. --Canley (talk) 02:22, 29 November 2019 (UTC)

@Canley: Interesting! Thanks. I'll not tweak any more locations then, if you're going to re-do the import. Thanks for working on this stuff! —Sam Wilson 03:24, 29 November 2019 (UTC)

Call for participation in a task-based online experiment

Dear Samwilson,

I hope you are doing good,

I am Kholoud, a researcher at King's College London, and I work on a project as part of my PhD research, in which I have developed a personalised recommender system that suggests Wikidata items for the editors based on their past edits. I am collaborating on this project with Elena Simperl and Miaojing Shi.

I am inviting you to a task-based study that will ask you to provide your judgments about the relevance of the items suggested by our system based on your previous edits. Participation is completely voluntary, and your cooperation will enable us to evaluate the accuracy of the recommender system in suggesting relevant items to you. We will analyse the results anonymised, and they will be published to a research venue.

The study will start in late January 2022 or early February 2022, and it should take no more than 30 minutes.

If you agree to participate in this study, please either contact me at kholoud.alghamdi@kcl.ac.uk or use this form https://docs.google.com/forms/d/e/1FAIpQLSees9WzFXR0Vl3mHLkZCaByeFHRrBy51kBca53euq9nt3XWog/viewform?usp=sf_link I will contact you with the link to start the study.

For more information about the study, please read this post: https://www.wikidata.org/wiki/User:Kholoudsaa In case you have further questions or require more information, don't hesitate to contact me through my mentioned email.

Thank you for considering taking part in this research.

Regards

Kholoudsaa (talk) 23:25, 13 January 2022 (UTC)

Please keep hotel and heritage building separate

regarding Fremantle Club building (Q66975227) Germartin1 (talk) 08:53, 25 September 2022 (UTC)

Sorry! Reverted. Sam Wilson 10:50, 25 September 2022 (UTC)

CiteTool

Hello! I just wanted to thank you for creating User:Samwilson/CiteTool.js. It has long been far too tedious to quickly generate citations, and this has made my efforts much easier. I don't know anything about scripts or programming, but I hope to see this tool (or something like it) improved in the future, e.g. possibly auto importing more metadata like publisher (P123) based on source url (e.g. NBC News (Q2877626) for a reference url beginning with www.nbcnews.com/ ). Do you know how widely this tool is used? Wikidata is so decentralized, arcane and idiosyncratic that it's hard to find quality tools like this (although it looks like your tool was recently added to Wikidata:Tools/Enhance user interface). Anyway, thanks again keep up the good work! -Animalparty (talk) 03:12, 23 October 2022 (UTC)

@Animalparty: Thanks! I didn't do the hard work, I just forked it and tidied it up a bit. I use it heaps! Thanks for alerting me to Wikidata:Tools/Enhance user interface; I've not looked there for ages, and there's some new stuff. As far as adding more and better metadata goes, I think the task for that belongs to Citoid, which should expose things like publisher Q-ID (although, I know there's long discussions about all that and am not really on top of all the arguments). To see how much a user script is used, a search for the script name gets us somewhat close (and seems better than the backlinks method, which relies on script users adding a wikilink comment, which not all do). Sam Wilson 04:29, 23 October 2022 (UTC)

@Animalparty: Thinking about this a bit more, I think perhaps what the tool should do is check the input URL against formatter URL (P1630) and then if it finds a match, follow Wikidata item of this property (P1629) from that property to the relevant item, and look there for publisher (P123). Does that sound correct? I think that wouldn't be too hard. @99of9: Entity Explosion does something like this doesn't it? Do you have any suggestions for how the CiteTool should approach this? Sam Wilson 06:40, 23 October 2022 (UTC)

Like I said, I don't know anything about the underlying programming. Prior to finding CiteTool I was (and often still am) using New Q5 Tool for pre-existing items, often with dummy values for existing items when the citation I wanted to add wasn't necessarily for birth or death (or if the item wasn't even a human). I don't know all the magic-science-logic that goes on under the hood, but it appears (at least to me) the New Q5 tool derives publisher (P123) from official website (P856), and provides fairly detailed, structured metadata. There is a problem with lots of unsourced statements (probably numbering in the high millions at least), and while bots can certainly play a role, making it super easy and more intuitive for humans to generate and add references would go a long way towards improving the quality and verifiability of data, especially for more obscure items that aren't already well codified in databases or bot friendly websites (e.g. a minor 19th-century city council member whose death date is only printed in a scanned book or newspaper). A tool that reduces the effort needed just to type the book or periodical title would go far, even if other metadata (date, title, page/column number) still needs to be hand entered. -Animalparty (talk) 07:11, 23 October 2022 (UTC)

@Animalparty: Yep, you're right, New Q5 is using official website (P856), checking a bunch of variations on the URL domain name to find a match (with and without 'www' etc.). That's probably a reasonable thing to do in CiteTool. I'll have a look. :-) Sam Wilson 07:45, 24 October 2022 (UTC)

I agree with you that this task sounds a bit similar to what Entity Explosion does for identifiers. Your proposed method makes sense. I dealt with a bunch of issues like www, but don't have much meta-advice, just deal with the problems as they come? --99of9 (talk) 05:27, 25 October 2022 (UTC)

@99of9: Thanks! Yes, good advice. :-) @Animalparty: I've had a go at adding support for just querying against official website (P856) and using that for publisher (P123). See what you think. I think one improvement is to check whether the item it finds has itself got a publisher (P123) and if it does, use that (e.g. in The Guardian (Q11148)publisher (P123)Guardian Media Group (Q2665980) we want the latter.). Sam Wilson 16:55, 25 October 2022 (UTC)

Would it be possible to make so links to The Guardian (Q11148)'s website adds published in (P1433) > TheGuardian.com (Q5614018)? --Trade (talk) 22:26, 24 November 2022 (UTC)

@Trade: I'm not quite clear of what rule we should institute for this. Because at the moment, CiteTool looks for a matching official website (P856) and if it finds a match then it uses it as publisher (P123). That doesn't make sense sometimes, certainly (see above) but I'm not sure when we can make the call to use published in (P1433). It might work, I think, to say that if the first item found with a matching official website (P856) has a publisher (P123) then that should be used for the reference's publisher (P123) and the first-found item be the published in (P1433) — but in your example that would result in a Guardian URL being given publisher (P123)Guardian Media Group (Q2665980) and published in (P1433)The Guardian (Q11148) (i.e. published in the newspaper and not the website). I think that problem comes about because both TheGuardian.com (Q5614018) and The Guardian (Q11148) have some of the same values for official website (P856). — Sam Wilson 22:47, 24 November 2022 (UTC)