Debian/Ubuntu Popularity Contest as a data source

[Eric]

Hi,

Here's another data source you might want to consider when you have
time. Ubuntu and Debian track the popularity of every package in their
repositiories, and publish the stats in a convenient format:

http://popcon.debian.org/
http://popcon.ubuntu.com/

The "vote" column -- number of active users, as opposed to total
number of installations -- is probably the most appropriate.

Pro: Counting the popularity of specific compilers and interpreters
avoids ambiguity with C/C++/C#, Lisp, Scheme, etc.

Con: This seems like it would count the number of people who have
installed a given compiler/interpreter. Actually, I think it also
counts packages that are a dependency of something else (e.g. CouchDB
depends on Erlang), which could inflate interpreted languages. But
that might be a useful indicator, too, if a language is popular and
useful enough to be used as the basis for other popular software.

[David Welton]

> Here's another data source you might want to consider when you have
> time. Ubuntu and Debian track the popularity of every package in their
> repositiories, and publish the stats in a convenient format:
>
> http://popcon.debian.org/
> http://popcon.ubuntu.com/

> The "vote" column -- number of active users, as opposed to total
> number of installations -- is probably the most appropriate.
>
> Pro: Counting the popularity of specific compilers and interpreters
> avoids ambiguity with C/C++/C#, Lisp, Scheme, etc.

Yeah, very precise results.

> Con: This seems like it would count the number of people who have
> installed a given compiler/interpreter. Actually, I think it also
> counts packages that are a dependency of something else (e.g. CouchDB
> depends on Erlang), which could inflate interpreted languages. But
> that might be a useful indicator, too, if a language is popular and
> useful enough to be used as the basis for other popular software.

I think the biggest 'con' is that we already have something of a free
software bias, between freshmeat, google code (which certainly doesn't
list code locked up in some bank), and I'm hesitant to add to that.
It is a good idea, however.

--
David N. Welton

http://www.welton.it/davidw/

http://www.dedasys.com/