-lgrassmath.a -lgrassstat.a (was Re: help [w/ s.semivar])

[James Darrell McCauley]

Shyh-Yuan Liu (syl at imt.ntou.edu.tw) writes:
>    I make s.semviar and display Error following:
> -------------------------------------
> ../meschach/LIB.sun4/meschach.a -lm 
> ld: Undefined symbol 
>    _memmove 
> *** Error code 2
> --------------------

Eric Kostello (kostello at math.ucla.edu) writes on 9 Sep 94:
>I had the same problem. You _have_ to follow the 
>instructions to replace machine.c and machine.h
>before you compile anything. I tried to do it in

yes, exactly. I really don't like including meschach
with s.semivar and I hope to wean it's dependence
on it.  It is only included to do the weighted
least squares fit of models to sample semivariograms.

I suspect that there's other code contained somewhere in GRASS that
might allow me get rid of meschach. If anyone knows of anything,
please let me know. I haven't had a chance to sit down and read all of
the source code lately ;) 

The dependence on MESCHACH is one of the items that has kept me
from transferring this to moon.cecer.army.mil from
pasture.ecn.purdue.edu (additionally, I'd like to have all of the
unfinished items in the current version completed. Then I can work on
the To-Do list in the file s.semivar/BUGS).

Since GRASS is mostly used as a research tool, I think that it would
be good if it had some sort of MATH/STAT library (well documented, of
course) that all tools could use... including matrix operations, maybe
some common CDF functions, etc. This is not something that I want to
put together by myself right now, but just an idea that I've had for a
couple of years (and the time seemed appropriate to share it).

I *believe* that the copyright on MESCHACH is such that it could
be included in GRASS without a problem, but it might be overkill.
Do many current GRASS progs (and any in the planning stages) 
require standard matrix operations? What might be some other
generic operations required? 

Irina and Helena have tried to document their interpolation library,
and if you've seen the manual for s.semivar, you'll see that I've
tried to document all C functions that it uses.  Hopefully sometime in
the future someone can take a look all of the GRASS code/documentation
and see where code is duplicated or where there's some generic
capabilities that other programmers would find useful. Sounds like a
big but useful job. Just food for thought...

--Darrell