5

A new tag appeared today . HCI is actually a subfield of computer science, and maybe psychology as well. Here is the new question

I did a brief scan of all of our tags (ugh) and found that few fields here have specific tags for sub-fields. But it seems that CS is the main exception with several, including and . Those sub-field tags have relatively few questions.

A further search for "human computer interaction" reveals 88 questions. But many (most?) of them aren't so tagged, nor even tagged with . And those can be found with a search without any tags (obviously), as the terminology is standard and in no way obscure.

As a "new-tag skeptic" I wonder if there is any consensus on the desirability of such sub-field tags. I replaced the new tag on today's question briefly, but rolled it back in favor of a discussion here first. It may be an edge case, actually. But the question itself is more general than the study of HCI and could be applied to at least CS and probably broader.

My personal view is that too many tags is a detriment here and that some users, especially new ones, feel that all keywords should be tags.

I'll note that many fields have recognized subfields, especially math and physics, but there don't seem to be sub-field tags for those.

So, the more general question, if answerable, is about how general or specific our tags should be. Likewise, should tags be a substitute for search.


It might also be the case that the CS sub-field tags are just an artifact that some of those fields are "hot" at the moment, making some people think they are especially important relative to others.

1
  • I'm not sure how to interpret votes on this question, especially down votes. Any enlightenment would be appreciated. This seems to be the case for several questions on meta here.
    – Buffy
    Jan 13 at 15:43

1 Answer 1

10

My 2 cents....

My personal view is that too many tags is a detriment here

I agree

So, the more general question, if answerable, is about how general or specific our tags should be.

My view is that tags should differentiate substantively different categories of questions.

  • If the question hinges on the differences between, say, measure theory and topology, then the question is not a good fit here. So, I would recommend against introducing tags for these subfields.
  • On the other hand, some subfields of biology do experiments with live mice or even humans. These introduce very different considerations, and so I wouldn't object to introducing tags for or .

A new tag appeared today .

At this point, we already have 466 tags with at least one question each, and a few hundred more that are synonyms. I think we have passed the point where users should simply create new tags: rather, new tags should be proposed and approved here on meta, just as we do for canonical questions. This is a good example: offhand, I am inclined to say that this field does not need its own tag, but if someone creates a proposal and makes a strong case that this field has its own peculiarities due to the way it takes data on humans using computers, I could be convinced.

12
  • 1
    Perhaps those proposing a new tag should also suggest at least the short form for the tag wiki. The wiki is seldom created for new tags.
    – Buffy
    Jan 13 at 19:04
  • Why is more tags bad? There's a tag limit, but I don't think we hit it frequently. Jan 13 at 21:30
  • 2
    @AzorAhai-him-, If tags have too few questions they aren't useful as categories. And if there are too many tags they are too hard to find and people will, then, just add more, making it worse. We have search for specific things and tags for categories of questions.
    – Buffy
    Jan 13 at 21:56
  • @Buffy If tags have too few questions they aren't useful as categories I don't agree, why do you say that? if there are too many tags they are too hard to find Are people scrolling through the tag list? I don't think so, I think most people type things into the tag box, so more tags, including synonyms enhances discoverability. Jan 13 at 21:58
  • @AzorAhai-him-, but a question can only have five tags at most. Imagine a situation where there was a tag for every keyword. I've found five limiting for a few questions, actually. I have no problem with synonyms, actually, but it takes an action to create them.
    – Buffy
    Jan 13 at 22:02
  • @Buffy Sure, but I was asking about your comment about too few questions - what's the minimum viable # of Qs, in your opinion? How does any topic start with "enough" Qs? Jan 13 at 22:29
  • @AzorAhai-him-, I don't have a minimum. The current average is 93 questions per tag. New tags have just one question of course. But if, after a year or so, a tag only has two or three questions, it isn't that useful on a general site like this one. Especially with the five tag limit.
    – Buffy
    Jan 13 at 22:53
  • 1
    There's a reasonable argument that tags are obsolete now that search is a solved problem, and we shouldn't bother with tags at all. As we've seen, it's very labor intensive and most people just search rather than scrolling through tags. But setting that aside, all the advantages of tags (being able to subscribe to or block certain tags; using tags to find interesting or relevant questions) break down if we have thousands of tags with four questions each. It's not just a matter of precision (giving each question good tags), but also of recall (giving tags all of the corresponding questions).
    – cag51 Mod
    Jan 13 at 23:58
  • A few more fields beyond biology that do research on live humans: psychology, sociology, economics, and likely more. The value of a possible tag human-subjects-research is not that it is a subfield of something (like biology) but that it identifies issues that pertain to multiple fields. Jan 14 at 10:40
  • 1
    Concermimg the last part of this answer, the creation of new tags: I don't think most creators of new tags do so intentionally, they simply type in a keyword they think should be a tag and this action makes the new tag spring into existence. Most of these new or newer users don't even know about meta. To prevent the creation of new tags from happening there would need to be a pop-up message when entering a "new" keyword in the tag bar.
    – Sursula
    Jan 15 at 7:30
  • 2
    There is a way to enable a tag warning: meta.stackexchange.com/questions/62591/…. Maybe we should request this, perhaps even requesting custom wording that new tags should probably be discussed on meta first.
    – cag51 Mod
    Jan 15 at 16:48
  • The tag warning is probably a good idea.
    – Buffy
    Jan 15 at 21:48

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .