[Linux-aus] Assistance in starting a new Open Source project

Avi Miller avi.miller at gmail.com
Sun Nov 17 12:35:26 EST 2013

Hey everyone,

This is probably going to sound like a ridiculous question, given the time I’ve spent in the Open Source community, but I’m looking for some advice on how to start what appears to be a brand-new Open Source development project, i.e. developing something that doesn’t seem to exist in any (open) form right now.

While I could probably bang away for months myself, I’m self-aware enough to know that my programming skills (or lack thereof) would probably make me abandon the project before it became useful. My personal skills also lie pretty much in the PHP-MySQL camp, and that may not be the most efficient toolchain for this project either.

So, because I’m not precious about this at all, and someone may know of a solution my google-fu hasn’t yet been able to find, here’s what I’m after:

"A podcast/media delivery statistics gathering tool”

Here’s how it [should|could] work:

1. All media on a website is prefixed with a tracking URL, i.e. if the media is http://server.com/path/to/music.mp3 then the CMS would write that out as http://tracking-server.com/track/server.com/path/to/music.mp3

2. The tracking server would track requests for the media file. This is where things become interesting because of HTTP 206 range requests - it would be great to track individual range requests and combine them so that we get a better view of listener/viewership.

3. The tracking server would either have an API or a pretty front-end with lovely graphs and statistics one could show a non-technical user to show them how their podcast/media delivery is going. This [should/could] be combined with GeoIP stuff, so we can get an idea of demographic information as well as player information, i.e. what device was used to listen to the content.

Along with this, I have some specific requirements for my own nefarious (i.e. volunteer) ends:

I want a hierarchical account/profile system where I can create a master account, say for a radio station, and then sub-accounts for each of the radio programs. The tracking URLs would then have to refer to the master/program combo, so for example, it may look like http://tracking-server.com/track/master/program/server.com/path/to/music.mp3

Then, users with sub-account access only get to see the statistics for the program(s) to which they have access, while users with master account access get to see the statistics for all programs that are children of the master, i.e. to get a view of the entire station.

I’ve found commercial operations that will provide this as a service, but even with a not-for-profit discount, to track the number of programs I need, it would cost me in excess of US$3500/year. And, I figure that if we can get this working, a number of small/medium community radio stations in Australia (and internationally) could strongly benefit and get off third-party podcast mechanisms.

Anyway, that’s my pitch. If you’re interested in this as a side-project or if you know of something that already does what I need, I’m all ears. I suspect something like this could possibly be created as a plugin to something Piwik, but I have a preference for a standalone solution.

Assuming I get some responses, my plan in mid-December/January is to apply for a funding grant to assist with the development of this. So, if you feel that you could contribute, but it may require some funding, that could possibly be arranged.

Thanks for reading!

More information about the linux-aus mailing list