Commons talk:WikiProject Postcards

Archives 2020, 2021, 2022, 2023, 2024

Data of the Adress Sides of Postcards?

The WikiProject Postcards currently is categorizing and capturing data of Postcards by many catgeories such as subject and location but mainly of the front. I thought about the question why we are not categorizing and collecting the data of the Address sides even more and documenting them on the project page. For example we could collect the name and place of the addresse throughː

1. Image Annotations
2. Transcription / Inscription Template (example) (possibly with https://ocr-test.wmcloud.org/)
3. Adding the address in Structured Data (example)
4. Georeferencing the addresses Street (example)
5. Categorizing the addresses Street (example)

In my opinion i think it would be quite interesting exploring or searching Postcards by Address e.g. through Wikimap. Let me know about your thoughts about that. CuratorOfThePast (talk) 19:26, 14 May 2024 (UTC)

I added the cordinates to a couple of postcards a while back. I could never decide if they should be for the sender or recivers address though and I don't think its possible to do both. I thought about doing it with structured data but adding addresses that way isn't intuitive. The best way would probably be to have seperate sender and reciever properties on the structured datas/Wikidatas end though. --Adamant1 (talk) 19:38, 14 May 2024 (UTC)

@Adamant1Alternatively we could also add Sender and Receiver in the Description. But this would be also rather suboptimal approach. So far i only encountered Postcards with receivers / addresses. There is one Property for addressee but only for Wikidata items not for text. So hypothetically the corresponding properties would be name of addressee, address of addressee and name of sender, address of sender? But street and housenumber would have to be modeled as qualifiers also point in time and location? CuratorOfThePast (talk) 20:22, 14 May 2024 (UTC)

I'll add the sender or recevier to the description if there's other information to add along with it. Its not the greatest way to it though and a lot of the file descriptions on postcards uploaded by GLAMs would make that hard. I've only seen a few postcards with the senders address myself, but I think having a uniform, easy way to add both addresses would be helpful for other types of post. That would be another benefit of doing it with structured data to. Then you could just search for "so and so address" and come up with results for postcards, postal covers, Etc. Etc. It wouldn't be as eaay or streamlined with descriptions or coordinate though. At least not on their own. --Adamant1 (talk) 21:19, 14 May 2024 (UTC)

@Adamant1In my opinion it would be imported to reference the address in some mean. Since the address probably would have changed or is defunct. I therefore somewhat prefer georeferencing. A combination of structured data with georeferencing would be good. So coordinate location should be allowed as qualifier. Do you have any suggestions to "suggested" properties? I would then maybe make a Wikidata Forum Thread and ask there also for opinions? CuratorOfThePast (talk) 21:40, 14 May 2024 (UTC)

Yes, Yes! I like the idea to put the address information in SDC. In the postcard collection of Graz they have done this (only town to town). You can see this on this website (with a cool map). - Maybe first we should create a new category like Category:Postcard was sent. So we see where we can add addresses. Also Category:Postcard was sent, but address unreadable for postcards with bad resolutions and so on. - I like to create a webtool at toolforge for search and better presentation of postcards (like a Postcard-Database). But I think current one key-thing is missing at commons. We need a way to say in SDC this is the other side of a postcards. So that we can easy connect the two sides of a postcard. Any ideas for that. --sk (talk) 05:19, 15 May 2024 (UTC)

Now that you mention it I was actually thinking of creating Category:Mailed postcards and Category:Unmailed postcards a while ago, but never got around to it. That way we could also have categories like Category:Mailed postcards with unreadable addresses and Mailed postcards of X country. It would be cool to also extend it to other mailed items, like postal cards. Although I think the proper term is technically "posted", but "posted postcards" just sounds clunky. Maybe we could create a "mailed: Yes/no" property or something on Wikidatas end along with it, whatever term we go with. --Adamant1 (talk) 05:45, 15 May 2024 (UTC)

I am also in favor creating a Category:Mailed postcards. Regarding the Postcard Database i had the idea for identifiers in commons for each postcards in the description. Something like PC000001F and PC000001B. And What about using the property inventory number and collection "WikiProject Postcards" in that context? But i do not know if that violates Wikidata?

The other way would be to model each postcard as a Wikidata item and link the Postcards with instance of to the specific postcard. But i dont know how to then interlink both sides to the postcards to the wikidata item and back? CuratorOfThePast (talk) 10:02, 15 May 2024 (UTC)

You don't need a identifier. Every image at commons has a unique identifier (the page ID). If you have an image like File:Hastings_County_Archives_HC02986_(21704100745).jpg you can use the "Page information" in the left menu. There you find the Page ID 76994413. You can also use this ID to get the image like request (see URL). And you can use this Page ID with the API. Like this API-request. - So we need only a good way to connect both images via SDC. - I know many postcards from museums are in wikidata as an object. This SPARQL find over 2800 postcards. But I think this is not a good plan. Maybe I am wrong. I think a postcard is a mass product and many postcards come in different variations. So it will be very difficult and many work. --sk (talk) 13:36, 15 May 2024 (UTC)

I only know of the properties is recto of (P2681), is verso of (P2682), image of backside (P7417), part of (P361) and has part(s) (P527) if this is helpful? To summarize we would needː

1. Property for Addresses on the backside of postcards (name of addressee, address of addressee and name of sender, address of sender)
2. Property for if the postcards was mailed?
3. New Category for Mailed postcards
4. Possibly a new property for linking the backside to the frontside?

CuratorOfThePast (talk) 19:51, 15 May 2024 (UTC)

@Adamant1@Stefan KühnI started a Wikidata Forum Discussion for the first property idea. CuratorOfThePast (talk) 20:03, 15 May 2024 (UTC)

I'm with Stefan Kühn about not having a Wikidata object for individual postcards. One way I think you could resolve the problem, and a few others along with, is by having an information template specifically for postcards. Something like Template:Photograph but with fields for the publisher, photographer, link to reserve image, addresses, postmarked date if any, Etc. Etc. I had thought about creating one myself a while back but I'm not really well versed in template creation. You might start a conversation about it someone to see if there's anyone who will create one for us though. --Adamant1 (talk) 02:13, 16 May 2024 (UTC)

@Adamant1@Stefan Kühn I made a property proposal for address of addressee on Wikidata. CuratorOfThePast (talk) 08:32, 16 May 2024 (UTC)

For the point if the postcards was mailed you we can use also the property date postally canceled (P9052)? CuratorOfThePast (talk) 19:18, 16 May 2024 (UTC)

That works. It's to bad your proposal is getting shot down, but I think we can just go their suggestions for now and then propose something else if/when those properties don't workout for whatever reason. Like I'm not really a big fan of using the "author" field for something like this but whatever. --Adamant1 (talk) 21:13, 16 May 2024 (UTC)

I think send and receive coordinates are better as structured data. Coordinates should only be placed with the information template to indicate the 'camera' (or artist) location and/or subject location, just as we would with a photograph. Pi.1415926535 (talk) 21:57, 16 May 2024 (UTC)

With the current modeling with Addressee and the use of qualifiers it is not possible to georeference the adress of sender and receiver properly? What we do about that? A second problem i encountered is that it is not possible to query these postcards or at least they wont show up if query for P1817 in general? (example), (example) (example of file)? CuratorOfThePast (talk) 10:12, 17 May 2024 (UTC)

@CuratorOfThePast: What do you think about having "Category:Postcards mailed from X location" and "Category:Postcards mailed to Y location" categories since the property proposal didn't work out? --Adamant1 (talk) 03:07, 25 June 2024 (UTC)

Yes, that is a good idea! It will probably be the case that there are not many postcards for "mailed from". But it gives the opportunity to later easier add structured data based of some property. CuratorOfThePast (talk) 10:09, 25 June 2024 (UTC)

Postcard intersection categories

Hi Adamant1, I see that you've been creating intersection categories like Category:Postcards of Holyoke, Massachusetts published by Springfield News Co. that combine publisher and location. I would recommend against creating these intersection categories, as they are unlikely to be useful to most users: someone looking for postcards of Holyoke is unlikely to care about which publisher it's from, while someone looking for examples of a certain publisher's postcards is unlikely to care about the location. For users in either situation, the intersection category means they have to look through multiple categories rather than one. Best, Pi.1415926535 (talk) 22:10, 22 July 2024 (UTC)

@Pi.1415926535: I tend to disagree. I certainly find them interesting myself and I think others do to since there's already thousands of similar categories. Ones that I didn't create BTW. Maybe you could argue it's unnecessary for small towns, but it certainly helps when it's a larger city where the main category has hundreds or thousands of images. But regardless, plenty of regional or national postcard publishers create postcard series for specific areas and it's interesting to sort them that way. @Stefan Kühn: I'm wondering what you think of this in particular since I think you were the who started doing it originally. As I've said, I think it's an interesting and helpful way to categorize postcards. I'm willing to defer to whatever other people think though as long as it's discussed first and there's a consensus to change the status quo. Otherwise this seems like a none issue. --Adamant1 (talk) 22:21, 22 July 2024 (UTC)

For locations that need subdivision, it's far better to divide by subject/location - common general subjects (buildings of... XYZ, bridges of... XYZ, etc), geographic subdivisions, and/or specific common subjects. That's a natural way of dividing that's likely to be useful and doesn't impede the ability of users to find something that they're looking for. The publisher is generally unrelated to the exact subject of the postcard, so it should never be the sole method of subdividing a geographic category.

I oppose subdivision by publisher (and, very similarly, by decade/year) unless the category is also entirely subdivided by subject/location (with no files in the main category), but I do not oppose it when the category is fully subdivided by subject/location. For example, Category:Postcards of Massachusetts by publisher is fine because every Massachusetts postcard is also divided by county, but Category:Postcards of Holyoke, Massachusetts published by Springfield News Co. is not good because Category:Postcards of Holyoke, Massachusetts is not fully divided by subject. I'm sure there will be some edge cases, but I think this is a good general rule. Pi.1415926535 (talk) 23:40, 22 July 2024 (UTC)

It should never be the sole method of subdividing a geographic category. I don't think anyone has ever claimed it should be or treated it that way. I certainly haven't and don't. There's plenty of times where postcards are already categorized other ways or where I've found multiple postcards of the same location and organized them that way after organizing them by publisher. Nothing stops anyone else from doing the same thing. Just like nothing stops anyone from doing it when the postcards are in Category:Postcards of Massachusetts by publisher or similar categories. Plenty of images in state level categories aren't in anything else. So the level of categorization doesn't have anything to do with it.

And just to point out one benefit to categories like Category:Postcards of Holyoke, Massachusetts published by Springfield News Co., there's been plenty of times where I've organized images of postcards from a specific publisher that way and then uploaded more postcards of that location specifically because I was interested in the publisher, but could really care less about the location itself. There's really no way to figure out which postcards we have for a specific location and publisher, or to "fill in the gaps" when we are missing postcards of a location from a specific publisher, if the images are all just dumped in state level categories. --Adamant1 (talk) 00:42, 23 July 2024 (UTC)

It should never be the sole method of subdividing a geographic category. What I mean by this is there should never be a situation like there currently is with Holyoke, where postcards of a location are subdivided by publisher and not by subject. If there are enough postcards of a city to subdivide them by publisher, all of them also need to be subdivided by subject in order to be findable by users. The consensus on project-wide discussions has been that intersection categories should be used sparingly and only when the parent categories are also subdivided in other ways.

I understand that the intersection categories are helpful for your specific use case, but I don't think that outweighs the downsides of them when the parent category is not also 100% subdivided by subject. Other methods, including searches, can find arbitrary intersections when needed: Special:Search/incategory:"Postcards_published_by_Springfield_News_Co." incategory:"Postcards_of_Chicopee,_Massachusetts" Pi.1415926535 (talk) 01:10, 23 July 2024 (UTC)

there should never be a situation like there currently is with Holyoke, where postcards of a location are subdivided by publisher and not by subject. The images already weren't subdivided by subject. So putting them in a publisher specific category has absolutely no bearing on that what-so-ever. It also doesn't stop anyone from organizing them by subject if they want to either. Your acting like the fact that the images not being categorized by subject has something to do with the "by publisher" category when it doesn't though. If anything it makes it more likely the images will be categorized by subject. Since again, that's often how I end up categorizing them that way. Otherwise I'd like to see some evidence that leaving images of postcards in "by state" categories makes any difference to that, but I've almost exclusively worked on organizing images of postcards over the last few years and I can tell you from organizing thousands of them that it makes zero difference and if anything just makes what you want more likely to happen.

The consensus on project-wide discussions has been that intersection categories should be used sparingly and only when the parent categories are also subdivided in other ways. Yeah, but the parent categories are already subdivide other ways in most, if not all, cases. Your just cherry picking the one example to suit your opinion. That could easily be solved by just organizing the images how you want like everyone else does and not trying to make this about something that isn't an issue. Also, this is sparingly. I certainly wouldn't advocate for doing it other situations outside of postcards and it's not even something that happens very often in that case because we don't even know who the publisher is most of the time to begin with. So your really over inflating this. I don't even do it myself unless there's enough images to justify it. Like Category:Postcards of Springfield, Massachusetts published by Springfield News Co. there was upwards of 70 images by the same publisher for the same location and the parent category had over 300 images. So I felt it was justified. I certainly wouldn't do that if it was only a few postcards in either one though. So it is being done sparingly. To the amount of 2,568 categories versus upwards of 500,000 images postcards. So it's really nothing.

I understand that the intersection categories are helpful for your specific use case. Who created Category:Postcards of Redding, California published by Brück & Sohn and how long ago was it created Pi.1415926535? I told you this is something that has been going on for years and that I wasn't the one who started doing it that way. And BTW all the images in that category are also categorized by subject. So your opinion is totally baseless. --Adamant1 (talk) 01:33, 23 July 2024 (UTC)

@Adamant1, @Pi.1415926535: People who are interested in postcards often have a special focus. Local historians only want to see postcards of their village (Example Webseite of a very good german page: Laupheim in Germany). Fans of bridges are only interested in postcards of bridges. And there are collectors who only collect postcards from one artist or one publisher (Therefor exists: Trader with big list of publishers). All subdivisions have their justification. - Here at Commons we often receive entire collections uploaded by publishers (Brück & Sohn or Curt Teich). These collections have to be pre-sorted somehow into states or federal states, because nobody systematically looks at 10,000 postcards one after the other in Commons. - So I also start the c:Category:Postcards published by Brück & Sohn by location and we have c:Category:Postcards published by Tichnor Brothers by state. And this is total ok. - But I agree with Pi.1415926535. It is bad for search queries. But I think this is a well known limitation of the category system. So for me it is ok to have intersection categories. But I also think we should focus more at Stucture Data for Commons (SDC). There we can insert, caption, location, publisher and so on. And we can search for all combinations of this, without the limitation of the categories. Also is SDC multilingual. And not all people who interested in postcards are speaking English. - Also within the category system we will never get a clean structure, because it is free to edit and to change for everybody. - @User:Pi.1415926535 at the moment we should go on with the intersection categories. I think it is acceptable. You can add your thematic categories beside the intersection categories. This should be no problem. --sk (talk) 13:21, 23 July 2024 (UTC)