Since Stack Overflow now has a large programmer population and a high degree of activity, it would likely be a good place for a language use index such as the TIOBE Programming Community Index. While this could be done via screen scraping the Tags page, or though the periodic database dumps, it seems that it would be a bit easier to be developed by someone with direct database access.

Additionally, a site such as Stack Overflow might be able to provide a more accurate picture of language since it reflects active issues by developers in an developer environment.

Update: One small update to this is that this wouldn't necessarily have be a real-time listing, it could also be a monthly listing like the TIOBE index which would also serve to smooth out any "bumps" in the actual use. Also, a couple responders have noted that there would be a skew towards .NET languages; however, this skewing may or may not be such a bad thing as it could also reflect the languages that are being used more in professional environments.

Update Two: Also, another thing to note as it might not be clear enough, but I'm not really referring to an industry wide standard index based upon the Stack Overflow data but rather an index that is more specific to Stack Overflow. Maybe a language/tag use trending chart is a better phrase to use?


I think that that should be done on the monthly data dumps as I can't see the benefit in having a real-time index. The variation in numbers of questions posted will give a very noisy picture. There may also be artificial peaks at the weekend (say) as people work on their pet projects at home, which could well be in the more obscure languages.

I agree with the other answers on the initial .NET bias of the site.

  • I've actually thought about doing the index myself as a personal project now that the data dumps are available; but since this is something that one of the StackOverflow developers might be able to do in an afternoon I thought I would put the idea out there first. – rjzii Jul 6 '09 at 13:13
  • The data dumps appear to be the best way to do this and now that the site is more mature there should be some interesting statistics in there. – rjzii Oct 7 '11 at 17:44

I think that given it's still (relatively) young age, StackOverflow is not a great indicator of the popularity of languages due to it's very large .NET-centric audience which is what the core audience consisted of when it went live.

Java and other languages like Ruby and Python and PHP are definitely very well represented in terms of questions and answers, but I think to use the site as a guage on the actual popularity of the languages may be a bit off.

  • StackOverflow has been around for about 10 months now so I don't think that it would still be heavily skewed towards .NET-centric languages at this point. While any sort of of count of the tags on a monthly basis would likely show a bias when it first came online, at this point the bias may or may not be gone this is something that would be revealed once the trends became available. – rjzii Jul 6 '09 at 13:11
  • 2
    I disagree. 10 months is still a relatively short time period in my opinion, even for the meteoric rise of the website. The primary userbase is .NET-centric, therefore most of the people they refer to the site will likely be .NET-centric, so on and so forth. – TheTXI Jul 6 '09 at 13:29
  • The .NET audience has been creating .NET questions and answers. The only way that other people will find the site is by searching Google and hitting a .NET question, until the other languages creep in. I think it will take a while before we can expect to see a balance on SO. – EndangeredMassa Jul 6 '09 at 18:15
  • 1
    @EndangeredMassa - Java is the second most popular language with over 14700 tagged items with languages such as JavaScript, PHP, and Phython all having over 7000 tagged items. At this point, I think we have enough information to at least make some interesting trending graphs. Also, it could be that StackOverflow is a bit more influenced by industry which, in my experience, has been using a lot of .NET languages lately. – rjzii Jul 7 '09 at 11:59

I would like to note that Perl has a low score on Tiobe, because the places where most Perl users ask questions, are not indexed by Tiobe.


I don't know if the results would be statistically sound. For example, I think most of Joel's (and perhaps Jeff's) audience on their own blogs are .NET developers, which skewed the original population (and therefore the oldest questions are heavily weighted in favor of .NET). Now that SO is public, this is lessened, at least a little, but any results still probably wouldn't be statistically valid.

  • 1
    Interesting point on the statistical validity of the metrics. Offhand they I think they would still be valid but you would have to put them in context. The top three language tags on SO are C#, Java, and C++ so I would argue that they are likely fairly well reflective of the current industry. Also, the TIOBE Index does show monthly trending which would be useful in letting us see if they are skewed and if so, if that skewing still applies. – rjzii Jul 6 '09 at 12:58

It would be interesting as far as the SO community goes, but it would have no real comparison to a "real" index until the community grows beyond it's strong .NET user base.

You must log in to answer this question.

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