If a Tag gets too proportionately large for a Stack Exchange site, it will need to be broken out on its own. Does anyone have ideas on when this should happen? Metrics? Algorithms?

Remember this is a discussion, not a feature request.

I think that any process of breakout would start with an Area51 site for the specific Tag, and then they could really see if there was enough interest.

  • 2
    Why would you think this?
    – ChrisF Mod
    Apr 7 '11 at 16:13
  • 3
    Can you give some past examples of what you mean? Apr 7 '11 at 16:13
  • Hopefully my answer makes some sense as to why I disagree, I suspect I may need to tighten up the wording later when my brain's slightly less frazzled =)
    – Rob
    Apr 7 '11 at 16:20
  • @bill, I don't know if any Tag on any SE site has reached a threshold yet, but a conversation on this came up and I got curious. Apr 7 '11 at 16:21
  • 5
    Instead of waiting for them to break out, can we just kick out [Android] now?
    – user1228
    Apr 7 '11 at 16:54

A site is defined when there are enough experts or professionals in the field interested in asking and answering on such a site.

So it's more important to consider the people who want the site than it is to consider the tags associated with the site.

Programming languages are reasonably ephemeral - Yes, C# is a big deal now, but it wasn't a decade ago. Programmers will learn the new tools and technologies associated with their trade, and over time that will change.

It's important that the site cater to and be centered around programming, not around a specific tag or interest within programming.

So the question is fundamentally flawed. There is no need to break a tag out of a site simply because it grows too large.


I suspect that using "tag size" as a way of determining "need" is probably a poor one.

If we assume that a given tag warrants its own site when it has the number of questions that superuser or serverfault has (which is a completley arbitary metric, but hey, it's a starting point!) then the tag should've been spawned from stackoverflow quite some time ago as serverfault has 72,790 questions and has 162,884 tagged questions. Regarding proportions, has been applied to nearly 11% of the questions on stackoverflow. If we based "break-out" decisions on a combination of volume and proportion, again it should've been spawned from stackoverflow quite some time ago. Even solely based on proportion, it'd be a good candidate.

All that said, what benefit would c-sharp.stackexchange.com give? I can't see one as a question regarding using WebMethods can be tagged , and all quite correctly. Should that question be posted to c-sharp.se or to stackoverflow?

Therefore, I refute the statement "If a Tag gets too proportionately large for a Stack Exchange site, it will need to be broken out on its own." (Proposals on area 51 really are the right approach.)

  • I assumed that by 'breakout', the process would start with an Area51 site. Edited into question. Apr 7 '11 at 16:22

I don't think tag metrics should count for too much. Rather user base and a definable realm of questions should help determine if there is a potential community to be built around a specific topic.

A good example of this is the new SharePoint Overflow that is on it's way to breaking out (The Area51 proposal has reached 100% commitment).

The SharePoint tags are reasonably sized and there is definitely some overlap between ASP.NET, C#, VB.NET, etc. tags. So why SharePoint and not C#? Again, it's not the tag(s) so much as the user base willing to form a community around a viable topic of limitable/definable scope. Tag size may correlate some, but I doubt it's going to be a great indicator of potential SE sites.

  • There's also the part where SharePoint is defined by things outside the scope of Stack Overflow as well as inside. There's a lot of management and design that simply will not fit on Stack Overflow, which I imagine contributed to the conception of the original SE 1.0 SharePoint Overflow.
    – Grace Note StaffMod
    Apr 7 '11 at 17:52
  • You don't think Tag metrics and user base size and question count have any correlation? Apr 7 '11 at 17:52
  • @Grace Note excellent point. I think this illustrates perfectly why tag size alone doesn't directly indicate a sustainable community (that would benefit being separate from StackOverflow). A reasonable tag size combined with a dedicated user base and a need (non programming questions) are much stronger indicators. Apr 7 '11 at 17:57
  • @Lance Roberts I didn't say there would be no correlation, I just said it wasn't a great indicator. Obviously several questions using a tag indicates a large user base and as stated a large user base is part of what makes an additional SE site possible (not necessarily desirable). But, as pointed out by Grace Note, there has to be more than just a lot of questions of a given topic to warrant breaking away from SO. Otherwise SO would eventually become a large collection of minor tags rather than the place for programming questions. Apr 7 '11 at 18:00

You must log in to answer this question.

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