Lines of Code and Man-power estimate for Maxima

Well, you can get a pretty direct cost estimate of the 1982 version of
Maxima by looking at the budgets of the Macsyma and Mathlab Consortium
efforts at MIT (and earlier work), or at the staffing of those projects.
You would of course have to remove costs related directly to supporting
users as opposed to developing software.

Off the top of my head, here's how I'd estimate it.  In the early period
(1963-1969), it looks as though there was an average of about 1.5 full-time
graduate student or research staffer (Engelman, Moses, Martin, Korsvold).
The program ramped up in 1969, and from 1969-1981 (12 years), on the average
(with much variation), the staff included about 2 professors (Moses, Martin,
Wang at various times), 3 full-time research and administrative staff (Jeff
Golden, Jonl White, Gosper, Ellen Golden, ... at various times), 5 pretty
much full-time graduate students (including contributions from outside the
Macsyma group proper e.g. from the MIT Plasma lab and outside users), and 1
part-time undergraduate.  This includes the work on the underlying Lisp
(Maclisp) in supprt of the project.  Not clear how to account for this staff
in traditional full-time person-years, since professors only devote about
1/2 their time to research, and all researchers (including professors and
graduate students) spend much of their time doing things other than
algorithm design and programming.  But as a very very rough estimate,
perhaps we could calculate (1969-1963+1)*1.5*1/2 + (1981-1969+1)*(2*1/2*1/2
+ 3 + 5*1/2 + 1*1/4) = 86.5 person-years.  Calculating at say $150k/py (that
is, current costs including overhead, not historical costs) gives $13m,
which is surprisingly close to the previous estimate.

RJF may have better estimates of all these numbers....

            -s

On Fri, Apr 23, 2010 at 20:29, Raymond Toy <toy.raymond at gmail.com> wrote:

> On 4/23/10 6:59 PM, Jaime Villate wrote:
> > On Fri, 2010-04-23 at 17:11 -0400, Boggess Rod wrote:
> >
> >> So I have to ask, why are you doing this estimate? Are you looking to
> >> estimate personal worth?
> >>
> > For no reason :)
> >
> > I just found interesting to apply sloccount to have an idea of the
> > number of code lines and programming languages  used in different
> > sections of Maxima (src, share, etc.). I was just too lazy to cut out
> > the part that I was trying to show (number of lines) and the message
> > went out also with the man-power and cost estimates which, in my view,
> > are meaningless. Sorry for the noise.
> >
> I thought it was fun, even though I don't believe the cost estimate.
>
> You can also get some information from
> http://www.ohloh.net/p/maxima/analyses/latest.  I didn't know there was
> some visual asic code and assembly code in maxima.
>
> It seems to me that ohloh used to include a cost estimate but perhaps my
> memory is failing.
>
> Ray
>
> _______________________________________________
> Maxima mailing list
> Maxima at math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima
>