Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: notesfiles
Path: utzoo!watmath!clyde!cbosgd!hplabs!hp-pcd!orstcs!nathan
From: nathan@orstcs.UUCP (nathan)
Newsgroups: net.micro.mac
Subject: Scientist's Helper posting complete
Message-ID: <21100035@orstcs.UUCP>
Date: Fri, 28-Mar-86 18:31:00 EST
Article-I.D.: orstcs.21100035
Posted: Fri Mar 28 18:31:00 1986
Date-Received: Wed, 2-Apr-86 01:39:43 EST
Organization: Oregon State University - Corvallis, OR
Lines: 52
Nf-ID: #N:orstcs:21100035:000:2653
Nf-From: orstcs!nathan    Mar 28 15:31:00 1986

Subject: Scientist's Helper posting completed.

I have just returned from vacation in sunny (?) Pullman, WA.  (Yes, I had
a wonderful time, and can say with authority that the best line in
Buckaroo Banzai is, "Laugh-a while you can-a, monkey-boy!".)

Parts 6 and 7 of the distribution have been posted to net.sources.mac.
The are two pieces of a binhex.  Strip off headers (and footers!!),
cat together, then xbin or Binhex, and PackIt_II to decompress.

J. Shulman asked if he could post SH to Delphi.  I can't seem to get
through via mail, so I say here: "go ahead, no skin off my nose!", or
more polite words to the same effect.  You're a credit to your nets, Jeff.

Now, a note to all who are interested in working with the source:
Scientist's Helper was written in Aztec C by an applied scientist
momentarily with Oregon State U. Dept. of Oceanography, William Menke.
(He'll be moving to Columbia U. soon, I don't know what department)
He is recent recipient of a presidential commendation or something for
his application of "Discrete Inverse Theory" to geophysical data
analysis.  He also wrote the book "Geophysical Data Analysis: Discrete
Inverse Theory, (Prentice-Hall).

This is relevant because, as anyone examining the code will note, he
is not a software engineer.  Part of the motivation for posting this
product is the hope that it will be "adopted" and we may see some
improvements.  Menke's attitude is that this is a quick-and-dirty
just to get some things done -- for which it works, but....

The first thing I hope to see done with it is porting to an environment
besides Aztec C.  I'm hoping the memory allocation difficulties will
go away by themselves then.  Also, I don't think it's really necessary
to include a 56K header file in front of each module.  Probably compile
speed could be improved, no?

The next thing desperately needed is a complete drop-in replacement
of the command parser, that can handle more than one arithmetic operation
per line.

The frequency portions of the program (Fourier xform, filter) are a bit
cumbersome too.  The FFT should work on a single column, and should be
"conservative" so the original data may be recovered.  (I propose
"mirroring" the data so that output is all real.)  An FIR filter would
be better than the provided Chebyshev one.

The Edit window needs work.  It should use 9-point font, not 10,
for those of us with floppies who have better uses
for disk space than too many fonts.  It should allow ranges of cells
to be selected for clipboard operations.  The scroll bars should move
proportionally to the actual list length.

Nathan C. Myers		orstcs!nathan	nathan@oregon-state