Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 158 Stack Exchange communities.

What is meta?
Here's how it works:
  1. Any Stack Exchange user can ask a question
  2. The community provides support, votes on ideas, and reports bugs
  3. Your voice helps shape the way Stack Exchange operates

I originally titled this as "How much coverage over the question's content should your tag choice revolve around?", and similar variants, all of which could be very confusing so I went with a less inquisitive title. I did look at the FAQ, but my particular concern was not addressed.

It is recommended that you use as many applicable tags as you can in order to increase the chances of people actually looking at your question. Sometimes, though, there are tags which are very much applicable, but only to certain portions of the question. I then think to myself whether I should add the tag, especially if it is a tag which seems to reduce the number of views.


I'll use an example question which I once pondered, but I never asked because I figured out that my currently implemented algorithm was as optimized as possible.

I need to remove a certain value from a delimited string. However, the string input will also include the delimiter at the beginning and end of the string. For example, I might have to remove 3;#Delta from the string ;#1;#Alpha;#2;#Beta;#3;#Delta;#. The algorithm would be different if the input string was at the beginning, end, or middle of the string. Can I optimize the following algorithm?


The obvious tags would be [c#] and [string-manipulation], maybe also [optimization] and [delimiter]. I could also include [sharepoint], because the whole ;# delimiter included both at the start and end is because I am using SharePoint. However, the actual question has nothing to do with SharePoint. It can be understood and answered quite correctly without any knowledge that SharePoint was in the situation. SharePoint is merely the environment that enforces why I am in the situation in the first place. So while [sharepoint] does have relevance to the situation and content of the question, it doesn't really apply to the essence of the question.

But maybe when I write the question, I might state "SharePoint" in order to just explain why I'm in the situation, but note that it is more specifically a C# question than even a C# and SharePoint question. Or maybe someone posts a comment asking "Why are you using such a stupid delimiter scheme?", and I would explain that it was because of SharePoint. If I include the [sharepoint] tag, then maybe I wouldn't have to answer that comment because it would be implied. I use a [sharepoint] example specifically because as far as I see, its presence does tend to reduce the number of people who look at the question. Perhaps it is because SharePoint has an equivalence between how evil it is and how useful it can be once you use it right, and maybe it might be because the people who address those questions are elsewhere. The reason is largely irrelevant because the example question is only marginally related to SharePoint. The tag would provide information as to the origin of the question, but it has little to do with the core of the code and the solution.


I write too much. I'm aware of that at times. But it feels wasteful to erase. Maybe I'm daft.

But even though I use a [sharepoint] example, I ponder about tag choice in general for all situations. Often times, you do have a very specific question that you'll be pounding your head against the 5 tag limit. But sometimes you aren't finding yourself at the 5 tag limit, and there are some tags which could apply to your question, but should they be used? Should you base your tag choice to hit all reaches of your question, no matter how remote? At what point does search coverage overtake the relevance of the tag? Should tags primarily be focused on the core of the question?

I'm not asking for a change in how people tag. I'm more curious as to people who have stumbled upon thinking what to tag their questions, if they run into this kind of situation, or people who do have some insight or recommendations to these things.

share|improve this question
up vote 1 down vote accepted

This is a really subjective topic since tags applied to a question are completely dependent on the question itself.

Generally speaking, however, I'll tag a question by starting with a list of tags (not limited to 5), generated like so:

  1. Classification/context tag(s), in descending order of importance. These are the most important tags that, if omitted or wrongly included, don't give the wanted solution. I.e., tagging with [perl] instead of [c#] when a C# solution is wanted is not going to end well.

  2. Direct subject tag(s), in descending order of importance. These tags identify the actual thing needing help, or being talked about. I.e., [string] means "I need help with how a string works" (usually accompanied by the [c++] tag), or "how do I use strings to solve problem XYZ?" These are normally tags that could be omitted and still allow someone to give an answer. For example, if someone wants to know the syntax of a for loop, they will still get a good answer as long as the correct language tag is on the question, even if [for-loop] is omitted.

Usually 5 tags is enough to give all that information. If I end up with more than 5, I'll start paring down the list by removing the least important tags from either (1) or (2) which has the most tags in the list, depending on the question context, of course. If a tag is specific to the situation (usually a subject tag), it should almost always be left in, possibly even at the expense of the least-significant context tag.

I've never run into a situation where either the context or subject couldn't be well described in 4 tags, leaving at least 1 tag to describe the other.

Meta-tags, like [sharepoint] in your example, should be removed, as they don't add any information, context, or value to the question. The question can be answered without any knowledge of SharePoint. It really isn't relevant, since if you changed the delimiter in your sample to something more normal, the point is completely moot.

That being said, if you feel the need to explain why something is the way it is ("I denormalized those 25 tables on purpose. Really!"), go ahead and mention what you need to in the question body, probably in an addendum at the end of the post which shouldn't be required reading for anyone to give an answer.

share|improve this answer
It perhaps is a very subjective topic, but your answer does cover the general approach, which is exactly what I was looking for. Thanks! – Grace Note Apr 26 '10 at 13:48

You must log in to answer this question.

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