That sounds like an excellent idea! If you would like some help, let me know and I would be glad to :) Cheers, - Tim On Thu, Jul 1, 2010 at 11:50 AM, Don Stewart wrote:

I think we need to standardise the presentation of this data, and provide a lib to access it. I'll think some more about it. Should be possible to automate it all now (that's mostly done), publish in a known location, and provide an API for queries.

...

If anyone wants to see the popularity of their particular package(s), dons has kindly left a .txt file with the information at the following URL: http://code.haskell.org/~dons/hackage/Jun-2010/http://code.haskell.org/%7Edons/hackage/Jun-2010/ popular.txt .

Of course, this being Haskell, I had to whip up a little script to get

inforichland: the

...

information for you.

It may not be the most elegant, but it only took a few minutes and it does the job. You'll find it attached :)

Cheers, - Tim

On Wed, Jun 30, 2010 at 5:08 PM, Don Stewart wrote:

Downloads and popular packages on Hackage for Q1 and Q2 this year.

http://donsbot.wordpress.com/2010/06/30/ popular-haskell-packages-q2-2010-report/

-- Don _______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe

On Thu, Jul 1, 2010 at 4:26 PM, Bas van Dijk wrote:

On Thu, Jul 1, 2010 at 6:50 PM, Don Stewart wrote:

...

I think we need to standardise the presentation of this data, and provide a lib to access it.

This, together with reverse dependencies[1], would be a nice feature for the upcoming hackage-server.

Regards,

Bas

[1] http://bifunctor.homelinux.net/~roel/hackage/packages/archive/revdeps-list.h...

It would be! Presently, in hackage-server, modules can sign up for an IO hook that executes whenever a package is downloaded. Here's what's left to do: 1. Make a data structure that has all of the download information in a useful form, with a feature that updates it (including csv backup) 2. Write a parser to migrate Apache logs to this form (didn't Galois already write a parser, actually? [1]) 3. Establish a URI from which Hackage dumbly serves the download data (as html, json, etc.) for anyone to study 4. Eventually, analyze the data server-side to add useful bits of information to the package index And there's also an upload hook, like hackage-scripts has, so reverse dependencies can be updated. —Matt Gruen [1] http://www.galois.com/blog/2009/03/23/one-million-haskell-downloads/