[Fwd: Re: Recommendation for Octave statistics functions]

[Michael D Godfrey]

Forwarded as you asked.

Michael

-------- Original Message --------
Subject: 	Re: Recommendation for Octave statistics functions
Date: 	Fri, 25 Jan 2008 11:35:46 -0800
From: 	Michael D Godfrey <godfrey at isl.stanford.edu>
To: 	John W. Eaton <jwe at bevo.che.wisc.edu>, bug at octave.org
References: 	<47951D24.9010803 at isl.stanford.edu> 
<18328.20144.286461.490386 at segfault.lan> 
<4798CB21.2070603 at isl.stanford.edu> 
<18329.37533.534448.924467 at segfault.lan>



>
> Yes, one of the biggest problems of Matlab is namespace control.  I see
> subfunctions, nested functions, private functions, and classes as each
> trying to solve some kind of namespace problem in a less than optimal
> way.  Inventing something new could help, or, more likely, just cause
> future compatibility issues.  If you are interested in discussing
> this, please bring it up on the maintainers list.
>   
I agree with this, and this problem affects more systems than Matlab.
Could I mention that what I know about namespace control, and more
generally, dictionary organization, is contained in a book that I,
David Hendry, Hans Hermans and Robbie Hessenberg wrote quite
a long time ago. I have to admit that I think the book contains some
useful ideas, mostly David's. I did a lot of the dictionary work in the
final version, based on David's original work. The book, in PDF, is at:
? /Machine Independent Organic Software Tools/ 
<http://qss.stanford.edu/%7Egodfrey/mint/mint.pdf>
That's the end of the advertising.

I know very little about how Octave resolves variable and function names,
so our ideas may not be directly relevant.

John: is bug at octave.org the right list for this? Is that what you meant by
"maintainer's list?

Michael