V10/cmd/movie/denker.mail
From research!vax135!jsd Fri Feb 27 21:15:49 1987
Message-Id: <8702280215.AA20817@vax135.UUCP>
Received: by nexus.UUCP (3.2/4.7)
id AA00714; Fri, 27 Feb 87 21:15:43 PST
Date: Fri, 27 Feb 87 21:15:43 PST
From: vax135!jsd (John Denker)
To: research!bwk
Dear Brian:
A status report.
I got your package and unbundled it.
Set up some directories.
The make went just fine.
The demo/example "isort.gen" failed, since the awk
we have around here has never heard of functions.
Can you send a less complicated example? Or a more complicated awk?
I made up the following example script, and tried it and subsets thereof.
It seemed to develop OK, but anim just sat there, apparently doing nothing.
////////////////////////////////////////
clear
t: text 3 4 Hi there
a: line 1 2 3 4
b: box 10 11 12 13
c: circle 5 6 2
////////////////////////////////////////
Perhaps you could provide a similarly trivial (correct)
example, and perhaps its developed version, to facilitate
testing, and to facilitate creating others by analogy.
Other constructive suggestions, to help the next customer:
1) Assume that the distribution is unbundled into a directory called XXX.
The makefile should move from XXX to LIB all the files that belong there,
and set the permissions (chmod a+rx ...). Perhaps this could be done
as a "make install" or something.
2) Also, the makefile should compile newer.c.
From research!vax135!bsw Tue Mar 3 01:46:42 1987
Message-Id: <8703030601.AA19940@vax135.UUCP>
Received: by nexus.UUCP (3.2/4.7)
id AA04984; Tue, 3 Mar 87 01:00:55 PST
Date: Tue, 3 Mar 87 01:00:55 PST
From: vax135!bsw (Ben S. Wittner)
To: research!bwk
Jon & Brian --
Here are a list of (bugs / suggestions / queries) concerning
your animation stuff:
1) (bug) Anim seems pretty confused about what to do if you proceed
after the end of the script. It leaves up the last
frame, and can't clear it. I conjecture that it has
forgotten about what's up, so it thinks the screen is blank,
and doesn't bother even to try clearing it.
2) (query / request) You indicated that the hooks exist
for implementing gray scales and colors.... How hard is that?
Is anybody at your end working on it? If not, do we stand
a chance of doing it here? It sure would be sweet.
3) My naive mind would have expected a "circle 1 1 1"
to be nicely inscribed in a "box 0 0 2 2", but that is not
the case. We noticed that boxes change their shape so as
to be a fixed fraction of the display window, but circles
are special. I am NOT convinced that this is a feature.
If this is really what you intended, it will require
some very delicate documentation. I suggest that the other
possibilities be provided also: circles that scale and/or
boxes that don't.
4) (Suggestion) Could you extend the language so that
the script can initialize some of the menu items? I find
it awkward that I must EACH TIME use the mouse to set up
"click x, click y, slow 32, step 1, again".
5) (Related suggestion) It seems particularly awkward
that getting to "slow 32" requires 5 mouse operations.
Entering numbers is a perennial unsolved problem in
mouseology, and I don't have a really good solution,
but perhaps you clever people can come up with something....
Possible (weak) constructive suggestion: a submenu
with a bunch of numbers on it, sorta like a scroll bar.
6) (Related suggestion) Several of the main menu items
are diametrically NOT self-documenting. For instance,
1-step is displayed only if you are NOT in 1-step mode;
xor .. .. .. .. NOT .. xor ..;
backward .. .. .. NOT .. backward ....
Possible constructive suggestion: on categories where
there are only two or three options, you could show ALL
of them, with N-1 of them SHADED OUT.
7) (Possible bug)
(( I have certain personal standards about turning in
"high-quality" bug reports, and this DOESN'T meet them.
I haven't done enough checking for dumb mistakes.
Feel free to ignore this, unless you can easily see
what the problem is....))
I had fdevelop blow up on me, on line 10164 of
a huge script. I looked at and near the "offending"
line, and didn't see anything unusual. I also extracted
the relevant lines with
head -10300 blot.s > xxyy
tail -300 xxyy > yy.s
and there were no problems with yy.s! Are there
any internal limits in fdevelop that were exceeded?
/usr/anim/lib/fdevelop: x value not a number
source line number 10164
/usr/anim/lib/fdevelop: x value not a number
source line number 10364
/usr/anim/lib/fdevelop: y value not a number
source line number 10364
/usr/anim/lib/fdevelop: unrecognized command
source line number 10564
/usr/anim/lib/fdevelop: garbage at end of line
source line number 10564
/usr/anim/lib/fdevelop: garbage at end of line
source line number 10763
/usr/anim/lib/fdevelop: string too long -- truncated
source line number 10962
/usr/anim/lib/fdevelop: string too long -- truncated
source line number 10962
/usr/anim/lib/fdevelop: string too long -- truncated
source line number 10962
/usr/anim/lib/fdevelop: string too long -- truncated
source line number 10962
/usr/anim/lib/fdevelop: garbage at end of line
source line number 10962
/usr/anim/lib/fdevelop: string too long -- truncated
source line number 11161
/usr/anim/lib/fdevelop: string too long -- truncated
source line number 11161
/usr/anim/lib/fdevelop: string too long -- truncated
source line number 11161
/usr/anim/lib/fdevelop: string too long -- truncated
source line number 11161
(( and on and on and on.....))
**********************************************************************
**********************************************************************
**********************************************************************
I also observed that anim hung up on big .i files.
Hmmm.....
Anyway, thanks again for sending us the stuff.
It has already taught us a lot about our simulations.
-- jsd
From research!vax135!jsd Tue Mar 3 19:48:42 1987
Message-Id: <8703040048.AA08820@vax135.UUCP>
Received: by cassandra.UUCP (4.12/4.7)
id AA04793; Tue, 3 Mar 87 19:45:15 est
Date: Tue, 3 Mar 87 19:45:15 est
From: vax135!jsd (John Denker)
To: research!bwk
Perhaps I should clarify what I mean by proceeding from the end
of the script:
For example, run your sorting demo (isort).
When it gets to the end, CLICK THE LEFT MOUSE BUTTON.
Anim will leave up the last display (and forget about it)
and restart the script, resulting in a fairly weird display.
It ends, of course, with the second result XORing the first
result into oblivion -- i.e. a blank screen. This is 100%
reproducible on our Suns -- we thought it was a strange feature
at first. If it doesn't reproduce on your systems, then it
is even stranger than I thought!
-- jsd