ohloh stats

[Guest lynx]

some more stats and linking:

http://www.ohloh.net/projects/10433?p=GemRB

 

Nice trend, huh.

 

The two files licensed under Artistic are:

gemrb/plugins/Core/snprintf.*

[Avenger]

The most interesting thing is that the pure code of GemRB worths almost 2M dollars

[Guest Guest]

I wonder how much WeiDU is worth.

 

-Galactygon

[Galactygon]

Oops.

 

-Galactygon

[Guest Guest]

A simple LOC ratio to gemrb would give a rough estimate.

[WizWom]

some more stats and linking:

http://www.ohloh.net/projects/10433?p=GemRB

 

Nice trend, huh.

 

The two files licensed under Artistic are:

gemrb/plugins/Core/snprintf.*

 

ER, 4000 lines of code a year is REALLY slow - That works out to just 2 lines of code an hour. I do that in a month. Easily. With testing.

 

Still even at 2 man-years, it's an impressive codebase.

[Guest lynx]

that's 4000 a person year, not gregorian year (31,5KLOC/year). Ohloh uses the simple algorithm http://en.wikipedia.org/wiki/COCOMO which isn't that appropriate for gemrb either.

 

If you check gcc, it comes just over 3KLOC/y in your calculation.

[WizWom]

140,000 lines of code should take much less than two man-years to implement, in almost any language. Not the 36 man years that ohloh states.

 

I don't care, really, whether some egghead came up with coeficients. I've worked with million line codebases which took small teams (8 people) a year for initial development and maintainence over a period of decades.

 

For instance, the original unix kernel, and initial tool set, about 200,000 lines of code, took 2 men about 8 months; but, of course, not everyone is a Kerningham or a Ritchie. Wozniak wrote the kernel for the Apple II monitor ROM overnight in assembler, about 8000 lines - and it went into production without change for 4 years. I'm nowhere near that, personally. I write a good solid 1000 lines a week, tested and clean.

[Guest lynx]

I don't care if you don't care about coefficients, but that estimate is calculated solely from them and KLOC. However wrong that may be.

 

Trying to manually calc the E gives me a different result.