On Wednesday 2009.01.28, at 04:48 , John Wilson wrote: > On Wed, Jan 28, 2009 at 9:48 AM, David Welton > <[email protected]> wrote: > [snip] > >> John Wilson says: >> >>> For languages which are of direct interest to this list I think the >>> best metric is mailing list usage (markmail.org is very good for >>> this) >> >> That's true for smaller languages, which probably includes a lot >> developed by people here (like my own Hecl, which is unfortunately >> significantly less popular than langpop.com despite being a lot more >> work on my part), but for larger ones, is going to be pretty much >> impossible to keep track of. I bet there are tons of Italian forums >> on PHP, for instance. Multiple that by languages, user groups, >> countries, and it's going to be impossible to keep track of. >> >> With anything though, you have to think about biases - lots of people >> like to talk about Haskell, for instance (the IRC channel is huge), >> but how much is it actually used in production? It's best to just >> throw lots of numbers out there and let people make up their own >> minds, is the conclusion I reached, and what I strive for with >> langpop.com.
In developing the enterprise version of Krugle and it's "management/ information dashboard", one of the areas that we tried to focus on was the trends and particularly the variations against the trends. >> What might be an interesting addition to langpop.com is a list of "up >> and comers", which could indeed use different, and potentially more >> accurate metrics. > > I'm quite interested in the popularity of JRuby rather than Ruby and > Jython rather than Python (and the popularity of Groovy and Scala as > well). > > With these languages and ones like them the mailing lists and IRC > channels are probably a very good source of data which indicate their > popularity. I think you would need to do some processing on the raw > data to get meaningful figures. I think a good indicator would be the > volume of user questions received. So you'd look to count only > messages from new and infrequently posting users and have some way of > counting threads as only one message. > > If it was possible to get a metric which was roughly "numbers of user > queries per month" I would imagine that would be quite a good > correlation with popularity. There would, of course, be a problem with > JRuby and Jython because they have a proportion of there users who are > new to the implementation but not new to the language so you would > expect to so a smaller number of messages because the "how to I turn > an int into a float" type of question would have been answered > elsewhere. If you're just looking at trends then thats true but trying to normalize those across communities (and, as David mentioned, across cultures, languages, etc.) that doesn't work so well. Also, one bias that highly connected, heavy online users like us have is that we really don't grasp just how large a portion of technical people don't regularly use the 'net or books. Yes, that frightens me to no end but it's one of those nasty facts. Take care, John --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "JVM Languages" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/jvm-languages?hl=en -~----------~----~----~----~------~----~------~--~---
