9

As per the latest blog post, what (specifically) are considered meta-tags? Jeff gave the examples , , , and Shog9's answer added and .

What other tags (that currently exist) should we discourage?

(Forgive me if this has already been addressed)

1

7 Answers 7

12

Enough with the "meta tag" witch-hunt. At some point, this is going to start doing more harm than good; heck, it's possible we're already there.

An inappropriate tag is a red flag for inappropriate questions

You think "discussion" is a bad tag because it's meta? Bah. [discussion] is bad because SO isn't a discussion site! Simply removing the tag doesn't do anything; there are still 400+ discussion questions hanging around. Either delete the questions with the tag, or leave it there as a reminder that cleanup is needed.

[code-golf] and [hidden-features] are great examples of this. Both identify things that, strictly-speaking, shouldn't exist on SO. They're not vague or confusing - both imply fairly strict standards for how the question will be asked and answered. I'm not a big fan of either sort of question, but if they're gonna be allowed to stay then we might as well classify them properly.

Discuss tags on their own merits

I'm having a hard time seeing a "meta-tag" classification for tags as any more useful than a "subjective" classification was for posts. Neither one helped to determine if the tag/post so labeled was useful or not, neither one was applied consistently, and if this discussion is any indication then neither one has an agreed-upon definition.

8
  • The definition is actually very simple: Does the tag tell you whether or not it's a subject you either have experience with or are specifically interested in? [code-golf] does, so that's OK. [hidden-features], not really. [homework], yes - there's a subset of people who took CS or who like to help students. But do [discussion], [offtopic], [rant], [knowledge], or the expunged [subjective] and [best-practices] narrow the playing field at all? No, none of them do, nobody's an expert on any of those things, and with the exception of Juliet's rebels, nobody's interested either.
    – Aarobot
    Aug 11, 2010 at 18:50
  • I do completely agree that lousy tags are a warning sign for a lousy question, but that's not always the case. When I went through several of the [untagged] questions a few days ago, it looked like about half of them could simply be tagged better and the other half were crap (which I voted to close). All the more reason to retag the good questions that happen to have bad tags; makes it easier to find the bad questions to close.
    – Aarobot
    Aug 11, 2010 at 18:51
  • 2
    @Aarobot: I'm having a bit of trouble seeing [discussion], [homework], and [code-golf] as being all that different. None say much about the topics themselves: you could take most questions in those categories and re-write them such that they asked effectively the same thing without requesting multiple / pedagogical / short answers. Their primary purpose then is finding / filtering questions that match a pattern - same as [subjective], but with perhaps better-defined usage.
    – Shog9
    Aug 11, 2010 at 20:35
  • 1
    I definitely agree that most of these tags deserve some additional scrutiny though. If it comes out that there's little actual commonality in real-world usage, that's probably a good indication that the tag is overly-ambiguous and should be removed.
    – Shog9
    Aug 11, 2010 at 20:37
  • Code golf is essentially an optimization problem, only it's optimizing for shortest code, not fastest running time. It tells you exactly what the question is about. [homework] is gray area, but since the CS 101 material tends to be pretty standard, I think it's fair to say that there is a specific group of people knowledgeable about that very subject. Seems to me that both are entirely valid for filtering/searching, but I can't say the same about [discussion] (except as an ignored tag, but it's not reliable for that).
    – Aarobot
    Aug 12, 2010 at 0:22
  • 2
    @Aarobot: if [code-golf] questions are about optimization, then they're among the most poorly-asked optimization questions on the site, the equivalent of someone asking, "What's the fastest way to do X (using any language, algorithm, or hardware)". They're not meant to have a definitive answer, or necessarily solve anyone's problem - they're just idle amusement, a game to play with like-minded programmers. [discussion] is the same thing: you're not necessarily looking to solve anything, you just want to talk about the problem, involve the other users in a conversation.
    – Shog9
    Aug 12, 2010 at 16:54
  • But code golf questions do have solutions. There may be many correct answers, but there is a best answer, and other answers can still be evaluated as right or wrong. And even though it may be basically irrelevant as a skill today, there were times when code size mattered. Sure, they're amusement, but they're amusement along a specific subject line, unlike [discussion] which doesn't tell you what the question is actually about. I don't like code golf either, but I urge you to consider the possibility that you're just lumping everything you dislike into the same category.
    – Aarobot
    Aug 12, 2010 at 19:26
  • 1
    @Aarobot: Actually, I like [homework]. Not a fan of [code-golf] and [hidden-features], but the consensus seems to be that they're allowed in limited circumstances, and so long as they are I don't see the point in removing their tags... We have made effective use of meta tags in the past: it's when they become too widely abused to control anymore that it's a problem (and I do not agree that "widely abused" is a synonym for "meta"). [subjective] got used for everything, and long ago passed the scale where it would have been possible to control it.
    – Shog9
    Aug 12, 2010 at 19:36
6

The tag is one.

3
  • 2
    homework, I am torn on, because although semi-meta, that tag does tell me with a fairly high probability what the question is likely to contain -- point being, CS101 is the same old CS101 in every university across the world. Aug 9, 2010 at 5:14
  • 3
    I think you being torn is significant, and the same is true about other meta-tags. "Subjective" being prime example. While it does not tell me anything directly about the content of the question, it tells me something about nature of the question. We have a communication system here in company where we facing the same problem. We ended up allowing meta tags, but distinguishing them (in our case all meta tags are required to start with a / prefix, dot would probably work as well).
    – Suma
    Aug 11, 2010 at 15:13
  • 2
    @Suma: In a closed system (private company), that's fine. On a global web site with 200,000 users, not so much. There are a zillion and one ways to describe the nature of a question, precisely zero of which actually help an expert to find and identify the subject matter. That's the difference between a Q&A site and a discussion board (or "communication system").
    – Aarobot
    Aug 11, 2010 at 18:55
5

The most egregious examples still in the system that I can see are and .

Lots of other one-offs in the top 100 questions alone, like:

That's just from about 2 minutes of digging!

Some of these might be considered important tags – although I'm not totally sure why – but they are definitively meta-tags.

4
  • 1
    how is [life] meta? It's VERY specific. This question is only for the living.
    – devinb
    Aug 11, 2010 at 16:15
  • @devinb: I disagree. Somebody who died of old age would have way more experience with that subject than anybody who's still alive, assuming they could answer the question, which they can't. So technically, the tag implies that the question is aimed at the deceased and should never receive any answers. If it does, then anybody who does answer is clearly BSing it.
    – Aarobot
    Aug 11, 2010 at 18:58
  • touche, good sir.
    – devinb
    Aug 11, 2010 at 19:07
  • 1
    [urgent-help-needed] - two such questions on stackoverflow. Aug 13, 2010 at 12:54
2

- it's a tag about tags!

1
  • But here on meta, meta tags are the only tags we can put on questions ;-P.
    – SamB
    Feb 25, 2012 at 19:33
1
2
  • 9
    I wouldn't say that [language-agnostic] is a meta-tag. If individual languages are tags then there needs to be a tag to say that the question is not specific to any language.
    – Aarobot
    Aug 11, 2010 at 15:30
  • 1
    The way to correctly tag as language-agnostic is to just add no language tag at all! Mar 16, 2012 at 16:35
1

– from the looks of it, this should actually be a synonym for ;)

0

Arguably, and are very often misused. There's a way to make it not meta-taggish, but more often than not it is. There lies, imho, the difficulty with meta-tags, the 1% cases where it's genuine.

You must log in to answer this question.

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