11
$\begingroup$

One of the most important questions to be settled is: How are we supposed to tag questions on this site? I happen to have some opinions on this matter, and I will use this meta post to lay out my 'vision'. Please provide alternatives by means of an answer to this question, or express your opinion by up- or downvoting this question. Most importantly, please suggest ways to improve this post! This could be done by improving my classification scheme, pointing out mistakes or anything else, really. Just leave a comment, or talk to me in chat.

Important things to keep in mind:

  • Tagging is important. It is crucial in keeping the site organized, allowing one to efficiently search the site, and helps those who feel the need to filter out content that they are not interested in.

  • The tags should be of use, and not just for this particular question. This means that a very specific tag (e.g. a specific date, name of a book) that is not likely to be used again (note that I am, here, talking about used several times again) is a bad idea. A 'pitfall' that I have seen several times already (although there are >20 questions on the site at the moment of writing) is tagging a question with the name of a scientist. I think this is generally a terrible idea, since one can hardly expect more than, say, 5 questions about a single person to come up. This means that this tag is not broad enough to be useful. Note that there may be a few exceptions to this general rule (Newton?, Einstein?), but this is something that should be discussed further at a later time, when we all have some more experience on the site.

  • The tags should be sufficiently descriptive that someone seriously involved in the general discipline which it applies to (e.g. mathematics, biology) will be able to tell what this question is about, and whether there is a reasonable chance that he/she will be able to understand the question (or answers) completely. This implies that just a single tag "mathematics" will not suffice: Tags need to specify different fields of research in order to be accurate enough to be useful.

Keeping all of this in mind, I propose the following rough classification scheme for good tags:

  • Discipline (e.g. mathematics, chemistry, computer-science, experimental-physics). These kind of speak for themselves. I think these are 100% necessary, so people can filter out questions on fields that don't interest them, if not for a 100 other reasons.

  • Subject matter (e.g. number-theory, quantum-mechanics, photosynthesis, machine-learning, organic-chemistry). The distinction between this category and the one above it is somewhat artificial and arbitrary, but certainly not harmful (in my opinion). Often, a clear-cut case can be made whether a tag should be a 'discipline-type tag' or a 'subject matter-type tag', but it is never essential that this should be possible, so there is no need to worry about possible inconsistencies.

  • Historical (and/or cultural) context (e.g. ancient-greece, renaissance, industrial-revolution). I just think it's a good idea to be able to separate questions based this criterion, both for searching and filtering purposes. Note, however, that the tag should be meaningful in the context of science. That is, a tag for a time and place should only exist if that time and place has characteristics that are (i) meaningful to the history of science and math, (ii) don't extend beyond that time and place, and (iii) apply to the entire time and place.

  • Type of question (e.g. reference-request, biographical-details, academia, examples). These will mostly be necessary to classify questions that are not of the standard format: "What was the historical development of idea X?" (very broadly speaking).

Note that this will, of course, not cover everything but hopefully it'll be sufficient to help along with the majority of the questions

$\endgroup$
  • 3
    $\begingroup$ I second this proposal and suggest to use discipline-agnostic for questions not specific to any discipline or academia for questions about the history of academia. $\endgroup$ – Wrzlprmft Oct 28 '14 at 23:05
  • $\begingroup$ @Wrzlprmft I don't think 'discipline-agnostic' is a good idea, because 'general-science' is already available for that, and much more obvious (in my opinion). $\endgroup$ – Danu Oct 28 '14 at 23:07
  • 1
    $\begingroup$ The problem with general-science is that it (technically) also comprises stuff that does belong a certain discipline or highly interdisciplinary aspects. The advantage of agnostic is that it sort-of explicitly excludes everything that does not belong to a given discipline. Compare also to the language-agnostic tag on Stack Overflow (think of a tag general-programming there) or system-agnostic on RPG. $\endgroup$ – Wrzlprmft Oct 28 '14 at 23:14
  • 1
    $\begingroup$ @Wrzlprmft hmm... I see your point. I don't like the tag name personally, but if it's an established tradition, I guess other people support it. Probably not without reason... ;) $\endgroup$ – Danu Oct 28 '14 at 23:16
  • $\begingroup$ Tagging by discipline is clearly necessary, and at a much more granular level than just (physics), (biology), etc. The average answerer will have a few fields which they can answer questions in and many fields which they would be totally unable to answer questions in, so we need a way to filter by discipline. I'm not sure discipline/topic is a useful distinction, but I agree with your proposal otherwise. $\endgroup$ – Jack M Oct 29 '14 at 0:16
  • $\begingroup$ For cultural context tags, I think it's important that the tags reflect historically meaningful contexts, not just (14th-century), (15th-century) and so on. That is, a tag for a time and place should only exist if that particular cultural context has characteristics that (i) apply to the entire context, but (ii) not beyond it, and obviously (iii) are relevant to the history of science and math. $\endgroup$ – Jack M Oct 29 '14 at 0:19
  • $\begingroup$ @JackM I incorporated that second suggestion in the post. $\endgroup$ – Danu Oct 29 '14 at 0:20
  • $\begingroup$ Should we open a meta thread where tags can be suggested and discussed? $\endgroup$ – Ali Caglayan Oct 29 '14 at 1:42
  • $\begingroup$ I guess we should. $\endgroup$ – VicAche Oct 29 '14 at 19:22
  • $\begingroup$ @Alizter I'm not quite sure. I think having a discussion in chat first would probably be a good idea, before posting on meta $\endgroup$ – Danu Oct 29 '14 at 19:26
  • $\begingroup$ Why is organic-chemistry a discipline, rather than a subject under chemistry? Compare quantum-mechanics as part of physics, or number-theory as part of mathematics. $\endgroup$ – Michael Weiss Nov 1 '14 at 18:56
  • $\begingroup$ I agree with Danu's initial reaction to discipline-agnostic. When I saw that tag, it never occurred to me it meant no-specific-discipline. I assumed it had something to do with agnosticism. The fact that other stacks have a similarly opaque tag is not an argument for following their bad example. $\endgroup$ – Michael Weiss Nov 1 '14 at 18:58
  • $\begingroup$ I think we should avoid overthinking things. Choosing tags after one writes a post should be quick process and not a Talmudic exercise. That's why all the tags mentioned in the initial post are good tags: I could tell immediately what was intended. $\endgroup$ – Michael Weiss Nov 1 '14 at 19:03
  • $\begingroup$ @MichaelWeiss the org. chem thing was, of corse, a mistake. It is now corrected. $\endgroup$ – Danu Nov 1 '14 at 23:25
5
$\begingroup$

The proposal seems in general good to me. However, I believe we should be rather careful regarding the "Type of questions" tags. They tend to be "meta tags" and these do not work so well in general. Some can be useful, but others are too unspecific to be useful.

To illustrate what I mean let me pick at one tag. (Sorry about that.)

What exactly should be tagged as and what is the point of doing this?

Plenty of question might ask for examples of something, plenty of others could be construed to ask for examples of something but OP will choose a different phrasing and not tag them as such.

Just the fact that something asks for examples does not seem like a distinctive or relevant feature of a question to me. Will there be anybody that cares specifically about all questions asking for examples. I have a hard time seeing this.

A good test for the quality of a tag is if it can work alone. I feel "examples" as sole tag is basically as good as no tag at all, thus it is not a good tag.

$\endgroup$
  • 1
    $\begingroup$ [examples] may not be a great tag, but I disagree that work-aloneness is a good test of a tag. You can search for combinations of tags (just put both in the search box). Much better to have [example] [physics] than have a combinatorial explosion of combination tags, like [example-physics], [example-chemistry], etc. $\endgroup$ – Michael Weiss Nov 1 '14 at 19:12
  • 1
    $\begingroup$ My objections to examples are seemingly even stronger than yours. Specifically, it seems to be intended for questions which are (surprise) asking for examples of something happening. Such questions are invariably list questions. The only difference with big list questions (which we've decided are off-topic) is that the list isn't yet "big", but the fundamental problems of subjectivity and uselessness are still basically the same. I'd not only say we shouldn't have the tag, but that we should be actively discouraging questions which would use it. $\endgroup$ – Logan M Nov 2 '14 at 4:56
  • $\begingroup$ @MichaelWeiss the point is if it does not convey any information alone it will not convey any additional information in combination with other tags. But, yes, I agree we should certainly not have example-subject. $\endgroup$ – quid Nov 2 '14 at 12:37
  • $\begingroup$ I am not arguing for an [examples] tag; I'm agnostic ;-) on that. I am arguing again the claim that "A good test for the quality of a tag is if it can work alone". [examples] may be a bad tag for practical reasons (as Logan argues), but a tag may be useful in combination with other tags. What's wrong with that? $\endgroup$ – Michael Weiss Nov 2 '14 at 14:19
  • $\begingroup$ @MichaelWeiss from the linked post "Meta-tags are actually a subset of a larger problem that I usually call dependent tags. These are tags that don’t say anything by themselves – you can’t tell what the question is about unless they’re paired with some other tag (or several of them). These tags are a problem because people don’t realize this and will often use that as the question’s only tag. " $\endgroup$ – quid Nov 2 '14 at 14:28
  • $\begingroup$ OK, good point. $\endgroup$ – Michael Weiss Nov 2 '14 at 14:38
4
$\begingroup$

Two points, neither exactly criticisms, just a question and a concern.

First, I'm not sure what is achieved by separating discipline from subject matter. What practical consequence does this have for deciding which tags to create? The distinction seems rather blurry to me. Is experimental physics its own discipline, or a part of physics? What about experimental biology vs. biology?

This leads to my concern. The division of science into separate disciplines is famously fluid over the course of history; much ink has been spilled in the literature on the changing meanings. Faraday, for example, did not really make a distinction between physics and chemistry; in the early modern period, everyone spoke of natural philosophy, not physics (look at the full title of Newton's Principia!); the term "scientist" wasn't invented until the late 19th C. etc. etc.

That said, the need for discipline/subject-matter tags is a no-brainer, but it's worth remembering that (for example) using the term "biology" to refer to Aristotle's On the History of Animals is anachronistic. These anachronisms are widely used in the history of science, by necessity, but also extensively debated.

We tend to think of the current taxonomy of sciences as natural and "God-given", but in fact it is largely a creation of the late 19th and early 20th C. While the discipline/subject tags are necessary, I think it's a mistake to think of them as "top-level". If anything, the division by period and region would be regarded as more fundamental by historians of science, I think.

This has some practical consequences. I see nothing wrong with tagging a question, say and nothing else -- no discipline tag at all. (Presumably the question is about methods of publication in the early modern period.) There is no more need for a tag than a question about astronomy needs a tag.

Second, we should be open to discipline/subject tags appropriate to other periods, e.g., . Or or : both Galileo and Kepler cast horoscopes (and not just for the money), and Newton spent more time on alchemy than on all the physics he's remembered for. A question tagged is probably off-topic, but one tagged probably not.

$\endgroup$
  • $\begingroup$ Good points. The separation between discipline and subject matter is indeed artificial, but removing it doesn't change anything in a practical sense. I will incorporate this remark into my post, though. $\endgroup$ – Danu Nov 2 '14 at 17:24
  • $\begingroup$ An advantage I see in having the discipline and subject tags somehow separate is that this could create some hierarchy of tags. On MO there are some dedicated toplevel tags (about 20 corresponding exactly to the categories used on arXiv a major sever for mathematical research articles). The guideline is to use (normally) at least one of those and then whatever in addition. We could do about this with the disciplines and maintain some list of those "major" tags. For the subject people could be more free then. $\endgroup$ – quid Nov 2 '14 at 17:45
  • 1
    $\begingroup$ I have problems with regarding discipline tags as "top-level", or at least more top-level than many other tags. I've added to my post to explain why. $\endgroup$ – Michael Weiss Nov 2 '14 at 22:49

You must log in to answer this question.

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