3 ms·

Programming Language and Development Tool Popularity by Actual Reported Usage

3 points by danielcer 17y ago

j_baker 17y ago
These are always fun to read, but it's always so difficult to believe them. There are so many different ways to measure them and different surveys always yield totally different results.
- davidw 17y ago
  Actually, most of my results at LangPop.com reinforce one another. Where they don't, explanations are usually simple and direct (Freshmeat and Ohloh.net don't have lots of non-open source stuff). So I think that generally these things do correspond to reality, albeit roughly.
carbon8 17y ago
Warning: this is only based on user-submitted, apparently unverified reports about a miniscule (<1000) number of sites.
- danielcer 17y ago
  Hm... no. I think you're just looking at the number of entries on the 'newest first' list. However, if you take a look at the number of entries for PHP alone there's thousands of individual websites. The results are based on self reports. However, assuming wiki spam is kept to a minimum, the results should be at least as credible as other heuristic measures, if not more so. (Disclaimer: I'm the guy who put together rankings page)
  - carbon8 17y ago
    OK, "18000+". Same thing in this context. Beyond that, there are a range of flaws with self reporting that are immediately apparent when you look at the entries and rankings. Most entries are grossly incomplete, and they are all very inconsistent. Some say "PHP" some just say "wordpress" others just say nothing more than something like "postgresql." The result is that "PHP" in the rankings doesn't actually represent PHP, even among the small number of sites that have been listed in your app. Furthermore, when I looked an hour ago, Erlang was ranked higher than Ruby, and now, apparently, "Django" represents 10% of the entries and "Ruby on Rails" is ranked much higher than "Wordpress," which is ranked just above "Rails," all of which are well above "JavaScript," "HTML" and "Linux." And so on. Normally I'd like to see any indie web app succeed and appreciate the time and effort involved in making one, but in this case the information is grossly skewed and claiming to represent something that it doesn't. People look at things like this and form opinions and beliefs on them.
    - danielcer 17y ago
      In the general list, Erlang is ranked at position 70 (Usage: 0.2%), while Ruby is at position 4 (12.7%). So, I'm not sure which list you were looking at where Erlang was ranked higher than Ruby. Also, entries like "HTML" & "Javascript" are ranked low since they tend to be deleted as being non-informative. All the criticisms that relate to self-report bias in the wiki and the completeness of entries are reasonable concerns. I'll try to change the text that introduces the rankings to address this.
davidw 17y ago
A nitpick (I emailed him, too). LangPop.com does not use recorded conversations. I only look at how many people are on channels. I don't want people thinking that they're being "spied on".
- danielcer 17y ago
  Fixed. Sorry, about that. I misread your method the first time around.

3 points by danielcer 17y ago

These are always fun to read, but it's always so difficult to believe them. There are so many different ways to measure them and different surveys always yield totally different results.

davidw 17y ago
Actually, most of my results at LangPop.com reinforce one another. Where they don't, explanations are usually simple and direct (Freshmeat and Ohloh.net don't have lots of non-open source stuff). So I think that generally these things do correspond to reality, albeit roughly.

carbon8 17y ago

Warning: this is only based on user-submitted, apparently unverified reports about a miniscule (<1000) number of sites.

danielcer 17y ago
Hm... no. I think you're just looking at the number of entries on the 'newest first' list. However, if you take a look at the number of entries for PHP alone there's thousands of individual websites. The results are based on self reports. However, assuming wiki spam is kept to a minimum, the results should be at least as credible as other heuristic measures, if not more so. (Disclaimer: I'm the guy who put together rankings page)
- carbon8 17y ago
  OK, "18000+". Same thing in this context. Beyond that, there are a range of flaws with self reporting that are immediately apparent when you look at the entries and rankings. Most entries are grossly incomplete, and they are all very inconsistent. Some say "PHP" some just say "wordpress" others just say nothing more than something like "postgresql." The result is that "PHP" in the rankings doesn't actually represent PHP, even among the small number of sites that have been listed in your app. Furthermore, when I looked an hour ago, Erlang was ranked higher than Ruby, and now, apparently, "Django" represents 10% of the entries and "Ruby on Rails" is ranked much higher than "Wordpress," which is ranked just above "Rails," all of which are well above "JavaScript," "HTML" and "Linux." And so on. Normally I'd like to see any indie web app succeed and appreciate the time and effort involved in making one, but in this case the information is grossly skewed and claiming to represent something that it doesn't. People look at things like this and form opinions and beliefs on them.
  - danielcer 17y ago
    In the general list, Erlang is ranked at position 70 (Usage: 0.2%), while Ruby is at position 4 (12.7%). So, I'm not sure which list you were looking at where Erlang was ranked higher than Ruby. Also, entries like "HTML" & "Javascript" are ranked low since they tend to be deleted as being non-informative. All the criticisms that relate to self-report bias in the wiki and the completeness of entries are reasonable concerns. I'll try to change the text that introduces the rankings to address this.

davidw 17y ago

A nitpick (I emailed him, too). LangPop.com does not use recorded conversations. I only look at how many people are on channels. I don't want people thinking that they're being "spied on".

danielcer 17y ago
Fixed. Sorry, about that. I misread your method the first time around.