Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!usc!zaphod.mps.ohio-state.edu!uakari.primate.wisc.edu!uflorida!haven!udel!mmdf
From: EVERHART@arisia.dnet.ge.com
Newsgroups: comp.sys.amiga
Subject: Re: AnalytiCalc woes
Message-ID: <11886@baldrick.udel.EDU>
Date: 22 Feb 90 20:03:29 GMT
Sender: mmdf@udel.EDU
Lines: 39

When I first wrote AnalytiCalc for Amiga, I had to size windows
on screen. (Fortran presents a fairly terrible interface to unearth
the actual screen size.) I assumed it would be used on an interlace
workbench (i.e., that windows in a 640 by 400 screen would be OK).
   If you run one of the older versions of AnalytiCalc on a noninterlace
screen, it will crash in weird ways; error 209 is one of the possible
outcomes. The newer versions ask first whether you have interlace or
non-interlace and size the screen accordingly. By the way, the code
currently assumes that you have the PCCHELP.HLP file in the
current directory; the keypad pictures won't work unless this is
true.
   The most recent version fixes a number of bugs and adds a bit of
speed. I FTP'd it to xanth.cs.odu.edu in the incoming/amiga directory
(files ana*.*) for any who are interested. 
   Error 209 is an AmigaDos error code.
   Incidentally, on a 512K Amiga, you should ALWAYS use the noninterlace
screen size for the thing; it's pretty tight and can run out of memory
and act weird if you try using the larger windows.
   You also should in a 512K machine set stack to the default 4000
to ensure THAT doesn't cause it to run out of memory. I'm not
certain what the error 209 is (my Amiga manuals are at home), but it
could also have to do with running out of memory. Extra stuff in
your limited memory WILL screw you up in that little space.
   The version I FTP'd had two sets of executables; one was for
1 meg or larger systems (I did not size it too stringently so
even 1 meg could be tight), and the other for less well endowed
ones. I finally added the original variability of size of the worksheet
into the sources. (It was designed in from the start, but I had
stripped it when converting to Microsoft Fortran 3.13 on an IBM PC
and had not put it back when converting THAT code for AmigaDos.)
The result is that the "bigger" AnalytiCalc can be compiled for
much larger memory use and need not rely on software paging as heavily.
   I don't recall needing any assigns to analyticalcexe: by the way;
I do no such when using it myself. There was some cruft I did to use
Iconx to fire it up off workbench; that works on the test diskette
I made. The only assign it needs is to dk:, and THAT only if the command
files are not in the default directory when you run the program.
Glenn Everhart
Everhart@Arisia.dnet.ge.com