Wikidata:Property proposal/Disused stations ID
Disused stations ID
editOriginally proposed at Wikidata:Property proposal/Transportation
Motivation
editThe Disused Stations website (Q114131866) is an outstanding online resource for information and photographs about former railway stations in the UK. We should already have items for almost all entries there, which should be straightforward to match by geographical location, but there may be some we are missing. It would be a useful site for Wikidata and Commons to systematically link to, and to confirm our completeness against; and a useful complement to our existing RailScot location ID (P10987) and Engineer's Line Reference (P10271) IDs. Jheald (talk) 13:32, 27 September 2022 (UTC)
Discussion
edit- Proposed. Jheald (talk) 13:32, 27 September 2022 (UTC)
- Notified participants of WikiProject Railways -- Jheald (talk) 13:36, 27 September 2022 (UTC)
- Notified participants of WikiProject UK and Ireland -- Jheald (talk) 13:37, 27 September 2022 (UTC)
- Support Railway stations are hard; disused railway stations doubly so ... towns frequently have had small handfuls of stations over the years. The Disused Stations website is well regarded - e.g. it is used as a source by Canmore (see, e.g. https://canmore.org.uk/site/339969/fort-augustus-railway-station ) - and can be expected to aid in the disambiguation and completion of the set of WD railway station items. --Tagishsimon (talk) 12:36, 28 September 2022 (UTC)
- Support. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:41, 28 September 2022 (UTC)
- Support. Possibly also useful alongside Railways Archive event ID (P2478). Thryduulf (talk) 18:16, 28 September 2022 (UTC)
- Comment Do we ever contact the owners of the websites before we do this? I can ask Nick Catford if you want. Secretlondon (not logged in and in a cafe). 85.255.233.86 11:36, 30 September 2022 (UTC)
- @Secretlondon: If you know Nick Catford then by all means go ahead. To be honest that isn't something I've done (though perhaps we should), in part I suppose because on the occasions I have been in touch with institutions I've pretty much had a brush-off. I did have a good discussion about potential duplicates with one institution, but even there when I asked for some data (eg info about IDs they had changed or updated, or about a new set of IDs they had just added) nobody there wanted to know. On another occasion another institution told me that giving me a list of updated IDs would be contrary to GDPR (???) Even with the most supportive institution, it's generally quicker to just extract the site from the web rather than to try to get a spreadsheet -- and it saves the contact there the trouble of trying to put it together. Certainly later we should be in touch with any anomalies or data glitches we have found (eg so far some grid reference linking anomalies at disused stations) -- but that's maybe more something for after we've had a serious stab at matching the data. Also once the data is in place, it can be a lot clearer for somebody to see what we're trying to do (and what we're not trying to do or replace), rather than for somebody to try to envisage it all in abstract up front. I do hope what we're doing is seen as above-board and fair-game, even if done unannounced (other than including a bespoke 'User Agent' tag in the html requests). A site like disused stations is a fantastic piece of work -- that's what makes it worth linking to. It seems to me that by adding links to wikidata we are doing much what a search engine would do -- not replacing the site (which contains far more information than wikidata will); but making it easier to find, and easier to generate groups of links to specific sub-sets of content within it. 'Disused stations' is a huge personal work by Nick Catford, and a huge contribution to the net, so I hope he will be okay with that. Jheald (talk) 12:56, 1 October 2022 (UTC)
- I think one issue is that our "ID" is someone else's website code, and not an actual ID. Secretlondon (talk) 17:53, 6 October 2022 (UTC)
- @Secretlondon: Maybe, but this is what we've done for a lot of IDs. For most (almost all?) WD properties what we call the "ID" is simply the shortest fragment of the url sufficient to identify an item there, that we can rebuild the full URL from sufficiently easily. In some cases that may be a numeric ID that their underlying database has minted, in other cases (as here) a specific short text slug. I'm not sure the difference signifies much -- still less if either way we also record the title given to the page with a subject named as (P1810) qualifier. Jheald (talk) 23:48, 6 October 2022 (UTC)
- I think one issue is that our "ID" is someone else's website code, and not an actual ID. Secretlondon (talk) 17:53, 6 October 2022 (UTC)
- @Secretlondon: If you know Nick Catford then by all means go ahead. To be honest that isn't something I've done (though perhaps we should), in part I suppose because on the occasions I have been in touch with institutions I've pretty much had a brush-off. I did have a good discussion about potential duplicates with one institution, but even there when I asked for some data (eg info about IDs they had changed or updated, or about a new set of IDs they had just added) nobody there wanted to know. On another occasion another institution told me that giving me a list of updated IDs would be contrary to GDPR (???) Even with the most supportive institution, it's generally quicker to just extract the site from the web rather than to try to get a spreadsheet -- and it saves the contact there the trouble of trying to put it together. Certainly later we should be in touch with any anomalies or data glitches we have found (eg so far some grid reference linking anomalies at disused stations) -- but that's maybe more something for after we've had a serious stab at matching the data. Also once the data is in place, it can be a lot clearer for somebody to see what we're trying to do (and what we're not trying to do or replace), rather than for somebody to try to envisage it all in abstract up front. I do hope what we're doing is seen as above-board and fair-game, even if done unannounced (other than including a bespoke 'User Agent' tag in the html requests). A site like disused stations is a fantastic piece of work -- that's what makes it worth linking to. It seems to me that by adding links to wikidata we are doing much what a search engine would do -- not replacing the site (which contains far more information than wikidata will); but making it easier to find, and easier to generate groups of links to specific sub-sets of content within it. 'Disused stations' is a huge personal work by Nick Catford, and a huge contribution to the net, so I hope he will be okay with that. Jheald (talk) 12:56, 1 October 2022 (UTC)