Re: Maple Vs Mathematica debugging / cost to write a system

Richard J. Fateman wrote: 
2 million lines of C++, hm.  I guess that would translate into
200,000 lines of Lisp.

take 150,000 lines of public domain code from the DOE Maxima system.

Don't use "average" programmers.

Suddenly it becomes much easier.

RJF

Note that the vast majority of people working for
software companies are not programmers. The programmers
are not even especially highly paid, compared to
marketing or sales people. I think WRI has far more
graphics designers than programmers.







parisse@xxxxxxxxxxxxxx wrote:

carlos@xxxxxxxxxxxx wrote:

Making schoonschip can be expensive.  Take a commercial-product
kernel of 2M C++ lines.  An average programmer can produce say 500
lines
of debugged C++ per year. Cost:  4000 man-years, or roughly $600
million.
Need to sell 4 million copies at $1500 each to recover the cost.


Didn't you mean 5000 lines per year?
That would divide the programming cost by 10. A commercial
project requires much more than just programmers, so maybe
500 millions $ for such a project is a correct estimation.




Some time (long) ago, when I was active as a Software Engineer(?),
software metrics told me that independent of language a
programmer could handle 40-50 lines a day. So we end up with
approx. 2000 lines of code a year.

I think this is rather pessimistic. With modern tools, using
proven technology, I think this numbers are irrelevant.
Building on available libraries software can be build far more
efficiently.


As RJF stated there is a bunch of tested code out there.
All we need to have an open source CAS is a Linus Thorvalds (linux)
and a Guido van Rossum (Python) to initiate the work. Maybe the
youngster Wolfram could be of any help. Certainly when he goes back in
his mind to 1979 when he was visiting Veltman and learned about
"schoonschip". Thinking about the ideals of what could have been the
benefits of a truly academic CAS, other than SMP and Mathematica.
No $-signs!

But how to cope with the 'not made here syndrome'?

As long as academics and workers for large companies have there
copies of Maple or Mma have for "free", there will be no big
motivation for development in the direction of a truly open CAS.

But maybe there is some light in the darkness. A few weeks ago I learned
about SAGE:
 http://modular.ucsd.edu/sage/index.html

An initiative of William Stein (with others, David Joyner, John Cremona
and many more). Not a general CAS, but in potention it could grow.
Shall we join the party?

Jaap
.

Relevant Pages

  • Re: Writing bulletproof code
    ... Good software design requires that you put error ... > the 'C' language and the libraries. ... It doesn't actually solve the problem that programmers are ... very few know when the cost will be negligible ...
    (comp.programming)
  • Re: Intro to Programming w/ Machine Language
    ... >> produced disk drives that are as fast as current technology allows ... And you're forgetting a crucial point here: there are manufacturing ... drives that affect the final cost. ... programmers get a "get out of jail free" card ...
    (comp.programming)
  • Re: Intro to Programming w/ Machine Language
    ... >> produced disk drives that are as fast as current technology allows ... And you're forgetting a crucial point here: there are manufacturing ... drives that affect the final cost. ... programmers get a "get out of jail free" card ...
    (alt.lang.asm)
  • Re: The economics of a slow but productive Ruby
    ... We programmers tend to fantasize that what we do is the most important ... involve capex and often represent a variable cost component in any ... the ones making the financial commitment. ...
    (comp.lang.ruby)
  • Re: What is the benefit to me of .NET as an end-user?
    ... even if moving to a new tool or framework were to double programmer ... 5% improvement in the productivity of the process. ... Since the cost of moving ... programmers using it that it is worth the cost, in the short run, when ...
    (borland.public.delphi.non-technical)