Subject: [Fwd: some Maxima questions -- 2nd attempt]
From: James Amundson
Date: 12 Feb 2003 10:42:36 -0600
I am forwarding this message on behalf of Tony Brocci, who is unable to
send directly to the maxima list.
--
James Amundson <amundson@fnal.gov>
Date: Tue, 11 Feb 2003 12:59:23 -0500 (EST)
From: R A Brocci <brocci at kaplmail>
Subject: some Maxima questions -- 2nd attempt
Jim Amundson:
First of all, I need to state that:
o Only a person with "root" privileges can install software
at our site.
I mention this simply because I don't know what else might
have been done after Maxima's "self install" completed.
o Maxima was installed under the XP OS, since it's the OS of
our (near) future, and I was unable to convince the system
types that I needed to have it installed under the NT OS,
as stated on your website.
o If you think, or if you prefer to handle, Maxima questions
via the phone, please let me know your number and a window
of time where dealing with questions would be convenient
for you, and I'll do just that.
o For what it's worth, all references to any Macsyma results
refer to (the former Macsyma, Inc's)
Unix Version 422
running on an SGI machine.
1. Your website indicated that I'd downloaded Version 5.6, but
the Maxima invocation says it's Version 5.5.
So, did I issue some erroneous click(s) on the download? (It
was done early this year.)
Should I download another version? In particular, is there
an XP version I should download?
2. I have been unable to capture a Maxima session via the
writefile( "some_name.extension" ) ;
command, in that the file is present but has zero size.
In addition, the use of the
closefile( "..." ) ;
command does not change this result.
Note: The same thing happens with the
stringout( "...", ... ) ;
command, which in turn means we haven't really been able to
exercise the
fortran( ... ) ;
command to the extent we'd like.
Note: This result seems strange, since the
batch( "..." ) ;
command works just fine.
Note: The writefile observation was confirmed by a colleague
who has Maxima installed under OS 2000 at home. (He says it
is also Version 5.5; i.e., the same version as here.)
Note: The
describe(writefile) ;
command responds with the
dsk, directory
arguments, but our attempts to interpret them with viable
writefile( "...", "..." ) ;
argument strings have been futile.
Note: According the mail archives, Tim Auton asked the same
question re the writefile command on 23 Apr 2001. With a hint
from Jim Amundson, I searched the archives for the writefile
string, and don't believe I missed the answer. (But perhaps I
did.)
3. Do you think this version of Maxima is capable of inverting an
N x N matrix, where N is > 10? I'm sure your answer is yes,
but:
For N = 24, a colleague let Maxima run over night, and the next
morning, it appeared from a screen back-scroll that Maxima had
"lost its way."
Note: I believe that N = 24 is a rather typical application for
this colleague.
This individual then used the
batch( "..." ) ;
command to import the Unix Macsyma version of the inverse of the
24 x 24 matrix, along with the terms of a 24 x 1 vector. With
these entries, he then issued the
B . V
command, where B is the inverse matrix and V is the vector. He
obtained the 24 x 1 null vector, which he knows to be wrong, and
which I could clearly see is wrong for just the
1st row of B . V
term. Any thoughts as to what might be happening here?
Note: For this colleague, it is typical for him to form a N x N
matrix (say A), invert it, and use the
stringout( "...", ... ) ;
command on the inverse in his "1st" session, and then import the
inverse and various vectors in subsequent sessions via the
batch( "..." ) ;
command.
In general, if you have any N versus inversion times on various
platforms, I'd appreciate seeing them. More precisely, I would
welcome any info on Maxima's matrix capability that you feel is
pertinent to the situation I've described above.
4. When I double click on your illustrative
describe(log) ; all
command, everything is fine. However, if I enter the
describe(log) ;
command, look at my options and then enter my answer, Maxima
does not respond unless I type
answer ;
which in turn leads to a diagnostic message associated with
the trailing semicolon before the next
(C...)
prompt appears.
Note: The "solution" to this situation is to enter
answer;
which doesn't seem to be in-sync with Maxima's philosophy of
not really caring where the terminating semicolon appears.
5. The word Macsyma appears in many places other than statements
regarding Maxima's origin; e.g., in some of the
describe(X) ;
commands I've issued.
I would think Maxima's origin is the only appropriate place
for Macsyma to appear. That is, whether one has never used
Macsyma, or is a long-time user like myself, Maxima's capa-
bility should stand by itself, in my opinion at least.
Or, are you attempting to distinguish between what came with
the Macsyma Prof. Schelter started with, and what your group
has done to enhance this version?
6. The
quit() ; or ctrl_C ctrl_G (without the ; character)
command does not shutdown the Maxima session; i.e., the window
doesn't disappear, and the "blinking cursor" appears after the
enter key is hit.
Is this result to be expected? If so, is it possible to bind
a site-specific version of
quit() ; or ctrl_C ctrl_G
to xmaxima? That is, and for example, a quit() function where
the Maxima screen does disappear?
Note: I mention this simply because we'd like to keep track of
who's using Maxima, and how often they're doing so. (I have
found that such info is absolutely invaluable when you have to
deal with the system types who seem to be always asking why we
need to retain package X.)
Note: I can handle the who's using issue via a (perl) script,
but I can't handle the frequency issue without having the
quit() ; or ctrl_C ctrl_G
command mean the same as your current
File->Exit
button.
Many thanks for your help!
Tony Brocci
KAPL, Inc
Mail Stop 123
PO Box 1072
Schenectady, NY 12301
518-395-6682 (voice) and -4422 (fax)
brocci@kapl.gov