Luis R. Rodriguez

Luis R. Rodriguez at

I finally am reviewing the COCOMO model [0] for cost of software as its developed. I'm looking at it since I am considering adding Coccinelle support for sloccount [1], which also includes cost estimates based on the COCOMO model. It would seem to be an easy fit for advancement of "ethical attributes". First step is to define the attributes that a free software license might contribute to. The attributes from the COCOMO model could be used, or new attributes could be defined to shape the algorithm. Ultimately I'd like to see a free software contribution index alone though.


Mike Linksvayer likes this.

Mike Linksvayer shared this.

Hi, IIUC the output of COCOMO is a projected lines of code, with various attributes (eg requirements volatility ... er looking at the wikipedia article, that isn't one of the attributes but those words are used in some of the actual ones) which can be assigned a financial cost based on labor market conditions where development happens. You're almost certainly aware that Wheeler and then others have estimated cost to create Linux and other FLOSS in proprietary environment based on sloccount, companion page

I suspect ethical attributes are mostly orthogonal to cost of production calculation such as COCOMO though cost estimation is one relatively easy input into gross estimation of free software contribution to various ethical attributes. Harder inputs would include measures of use and impact on proprietary competition.

What are the ethical attributes you have in mind that free software contributes to? I think listing those out (I and probably others also have ideas) would be a reasonable next step. Then for each one, analyze whether estimated cost of free software is a useful input into characterizing the contribution of free software to furthering the ethical attribute, and what other inputs would help.

I did re-skim but don't see specific ethical attributes called out.

My ur attributes would be freedom, equality, and security, but those need to be broken down to get to something measurable. Equality is probably the easiest as there are existing quantitative measures that it should be possible to estimate and project free software impact on.

Mike Linksvayer at 2015-06-04T16:50:06Z

"various attributes [as modifiers]"

Mike Linksvayer at 2015-06-04T16:52:16Z