Open Source Bill

[paulhildebrandt]

I've been thinking about funding Open Source software for a while. I think something (a program?) that would generate a bill would be very handy for visibility and informing a company. It would list OS software used, suggested donation based on use, and how to pay each line item of the bill. This way a company that wants to do the right thing can now see a clearer path.

[freakboy3742]

Sounds like a great idea to me.

As always, the devil will be in the detail :-) How are you proposing to gather the data?

My immediate thought would be a website where you answer a survey, with some answers possibly given by automated tooling you can run in a given project. For example, write a tool that does a pip freeze in a python project, and sends those details to the server to populate the answers of "what projects do I use".

But - a well coordinated publicity campaign about the problem with this website/tool as a focal point could be just the thing to get the fundraising ball rolling.

[FRidh]

I like this idea also very much. As @freakboy3742 mentioned, how do you gather the data, that is

• which software is used?
• and what should the suggested amount of donation per software be?

Determining which software is used can be done using package managers, especially with the Nix package manager and Nix Shell.

Say for now that you have a total amount you want to donate. You could simply divide the amount over each item equally, which would imply the lower building blocks (compiler) would receive a higher amount in total donations since they would receive more donations. Or you would use indicators based on amount of commits or lines of code as a key. Obviously, indicators are not perfect either (see e.g. in academia), and neither are these two specific indicators.

You could also try to generate a complete graph of all the software out there. Of course, this would only be an approximation. Such graph could help in determining the relevant importance of each software.

Anyway, in the end it all boils down to putting a value to something, which is hard.

. Both can be done with the

With the latter graph By comparing the former with the latter, and

[paulhildebrandt]

Okay cool, I didn't get shouted down and told I was crazy.  This is really good feedback.  I will expand the idea and address the feedback tonight.

[nayafia]

Just saw this, could be good inspiration for data gathering: http://popcon.debian.org/ They track Debian package use through an automatic survey run each week on users who have installed the popularity-contest package.

I'd start with listing OSS software used. The donation part is harder b/c not every project needs or wants donations, either, so hard to automate without understanding needs of each project. Maybe quantifying the actual usage of each project in their own company software would help prioritize the project list.