This particular case is a bit confusing. We've been using the "long read" tag for... articles that take a long time to read (personally, I usually use it if the article's over 3000 words or so)....
This particular case is a bit confusing. We've been using the "long read" tag for... articles that take a long time to read (personally, I usually use it if the article's over 3000 words or so). However, the post you linked to is from the site named "longreads", so it was tagged with that, which isn't quite the same.
I agree with you that we shouldn't tag posts with the domain they're from in general though (and I would have removed that "longreads" tag, but I'll leave it now so this post makes a little more sense). Really, we probably shouldn't need a "long read" tag either and could just do that based on word count, but the word count system is still fairly new and incomplete (it doesn't work on all sites).
We do have a very minimal tag-rewriting already, which (I think) only rewrites "spoilers" to "spoiler" so far. There's also a merge request that enables some autocomplete behavior for tags, which will help a lot. It's most of the way there but still has a few issues that I need to work through.
I've sometimes tagged video posts with the youtube author. For some youtubers (Wendover, CGPGrey, Veritasium, Contra, etc) I think it makes a lot of sense. This is useful I think, and I could see...
I've sometimes tagged video posts with the youtube author. For some youtubers (Wendover, CGPGrey, Veritasium, Contra, etc) I think it makes a lot of sense. This is useful I think, and I could see the argument for doing it with domain names (sometimes).
Yeah, it definitely makes sense to label the creator/channel of YouTube videos, either in the title or a tag. Again, it's something we can do more properly/officially, but that's not there yet at...
Yeah, it definitely makes sense to label the creator/channel of YouTube videos, either in the title or a tag. Again, it's something we can do more properly/officially, but that's not there yet at this point, so it's best to just use the title/tags for now.
Tags can't contain periods, and we shouldn't use tags for domains anyway. We already have the domain and it's already being displayed on every post, putting it in the tags just duplicates it. The...
Tags can't contain periods, and we shouldn't use tags for domains anyway. We already have the domain and it's already being displayed on every post, putting it in the tags just duplicates it. The search system does need to be able to search domains though (which it doesn't do currently).
I would say that yes, you shouldn't be adding tags like "road and track", unless they add information that's not redundant with the domain itself. As mentioned in other comments, something like...
I would say that yes, you shouldn't be adding tags like "road and track", unless they add information that's not redundant with the domain itself. As mentioned in other comments, something like the YouTube channel name isn't redundant, it's separate information. But a "road and track" tag when the domain is "roadandtrack.com" doesn't really add anything that isn't already there.
Being able to click the domains to see all posts from that domain would definitely be nice. Again, this is mostly just dependent on the search system having support for it, and then we can probably just link it to a search results page for that domain.
The ~creative situation is a little more tricky because the domain isn't shown in the listing any more, but I generally don't think it's very important. The specific site hosting the image doesn't matter much, and the information is still available via the site's icon (assuming you can recognize the imgur logo) or mousing over the link (assuming you're using a mouse). I don't think an "imgur" tag would really add much.
It is a bit confusing, yes. And I agree with you that this particular case it makes sense to remove the tag altogether because we have the information elsewhere in the post (which corroborates the...
It is a bit confusing, yes. And I agree with you that this particular case it makes sense to remove the tag altogether because we have the information elsewhere in the post (which corroborates the argument I made in my OP). Would you think it reasonable to retire the tag when the word count feature is more robust?
I had no idea about the spoilers -> spoiler rewrite, that's cool! That's exactly what I had in mind in my post and it's really encouraging to see you thought of it already. :D
First, I'm not trying to pick on anyone with this post, or single out anyone as doing something bad, but it's an example I saw and I wanted to have a discussion about it so we can all make the...
First, I'm not trying to pick on anyone with this post, or single out anyone as doing something bad, but it's an example I saw and I wanted to have a discussion about it so we can all make the site better going forward.
This post came up and I'm wondering about how we do tagging here. I've helped run larger sites that have tag collections (books, in this case, not links / discussion posts), and I learned a good deal there. One thing I learned is to not have tags for something you can search for otherwise. For books, this is things like authors, titles, publisher, that kind of thing. For submissions here, it's the domain, ideally one day the submitter, that kind of thing.
Furthermore, the longread tag is a duplicate of the long_read tag. I have a lot of experience with deduplicating tags like this, and one thing that worked really well was a silent tag rewriting system: something that would, on input, look for tags that have been marked as duplicates and then change them to the canonical version in the database and on display. This addressed common typos but also helped us standardize on one specific tag for something that could be referenced in a number of different ways. It's important to not interrupt the flow of content submission with this, but document it so the user knows what's happening. If they submit something with tag foo and it shows up in their post as food, we want to be sure they know why so they don't try to change the tags.
Again, not meaning to pick on anyone, but this is something I have some experience with and would like some feedback on.
I agree that we shouldn't create redundant tags: We don't need to tag something as "movies" when it's posted to ~movies. We don't need to tag something as "new york times" when the domain shows...
I agree that we shouldn't create redundant tags:
We don't need to tag something as "movies" when it's posted to ~movies.
We don't need to tag something as "new york times" when the domain shows "www.newyorktimes.com".
This particular case is a bit confusing. We've been using the "long read" tag for... articles that take a long time to read (personally, I usually use it if the article's over 3000 words or so). However, the post you linked to is from the site named "longreads", so it was tagged with that, which isn't quite the same.
I agree with you that we shouldn't tag posts with the domain they're from in general though (and I would have removed that "longreads" tag, but I'll leave it now so this post makes a little more sense). Really, we probably shouldn't need a "long read" tag either and could just do that based on word count, but the word count system is still fairly new and incomplete (it doesn't work on all sites).
We do have a very minimal tag-rewriting already, which (I think) only rewrites "spoilers" to "spoiler" so far. There's also a merge request that enables some autocomplete behavior for tags, which will help a lot. It's most of the way there but still has a few issues that I need to work through.
I've sometimes tagged video posts with the youtube author. For some youtubers (Wendover, CGPGrey, Veritasium, Contra, etc) I think it makes a lot of sense. This is useful I think, and I could see the argument for doing it with domain names (sometimes).
Yeah, it definitely makes sense to label the creator/channel of YouTube videos, either in the title or a tag. Again, it's something we can do more properly/officially, but that's not there yet at this point, so it's best to just use the title/tags for now.
Maybe include the TLD in the tag?
longreads.comlooks dissimilar enough tolong readto avoid confusion.Tags can't contain periods, and we shouldn't use tags for domains anyway. We already have the domain and it's already being displayed on every post, putting it in the tags just duplicates it. The search system does need to be able to search domains though (which it doesn't do currently).
I would say that yes, you shouldn't be adding tags like "road and track", unless they add information that's not redundant with the domain itself. As mentioned in other comments, something like the YouTube channel name isn't redundant, it's separate information. But a "road and track" tag when the domain is "roadandtrack.com" doesn't really add anything that isn't already there.
Being able to click the domains to see all posts from that domain would definitely be nice. Again, this is mostly just dependent on the search system having support for it, and then we can probably just link it to a search results page for that domain.
The ~creative situation is a little more tricky because the domain isn't shown in the listing any more, but I generally don't think it's very important. The specific site hosting the image doesn't matter much, and the information is still available via the site's icon (assuming you can recognize the imgur logo) or mousing over the link (assuming you're using a mouse). I don't think an "imgur" tag would really add much.
It is a bit confusing, yes. And I agree with you that this particular case it makes sense to remove the tag altogether because we have the information elsewhere in the post (which corroborates the argument I made in my OP). Would you think it reasonable to retire the tag when the word count feature is more robust?
I had no idea about the spoilers -> spoiler rewrite, that's cool! That's exactly what I had in mind in my post and it's really encouraging to see you thought of it already. :D
First, I'm not trying to pick on anyone with this post, or single out anyone as doing something bad, but it's an example I saw and I wanted to have a discussion about it so we can all make the site better going forward.
This post came up and I'm wondering about how we do tagging here. I've helped run larger sites that have tag collections (books, in this case, not links / discussion posts), and I learned a good deal there. One thing I learned is to not have tags for something you can search for otherwise. For books, this is things like authors, titles, publisher, that kind of thing. For submissions here, it's the domain, ideally one day the submitter, that kind of thing.
Furthermore, the longread tag is a duplicate of the long_read tag. I have a lot of experience with deduplicating tags like this, and one thing that worked really well was a silent tag rewriting system: something that would, on input, look for tags that have been marked as duplicates and then change them to the canonical version in the database and on display. This addressed common typos but also helped us standardize on one specific tag for something that could be referenced in a number of different ways. It's important to not interrupt the flow of content submission with this, but document it so the user knows what's happening. If they submit something with tag foo and it shows up in their post as food, we want to be sure they know why so they don't try to change the tags.
Again, not meaning to pick on anyone, but this is something I have some experience with and would like some feedback on.
It covers some useful things, but I'm proposing something a bit beyond that.
I agree that we shouldn't create redundant tags:
We don't need to tag something as "movies" when it's posted to ~movies.
We don't need to tag something as "new york times" when the domain shows "www.newyorktimes.com".