Trebuchet calculation/optimization in Perl?

[usenet]

OK, this is probably about as "misc" as it gets...

I'm interested in a Perl script to calculate and optimize a trebuchet
design (the siege engine, not the font).

As usual, I begin my project with an investigation of prior art (as all
'lazy' programmers should do). While I can find such programs in other
languages (such as FORTRAN), I cannot find such a body of work in Perl.
I half-expected to find a module on CPAN, but alas...

If anyone if familiar with such a project, I would appreciate a
reference to it.

Thanks!

 

[Bob Walton]

OK, this is probably about as "misc" as it gets...

I'm interested in a Perl script to calculate and optimize a trebuchet
design (the siege engine, not the font).

As usual, I begin my project with an investigation of prior art (as all
'lazy' programmers should do). While I can find such programs in other
languages (such as FORTRAN), I cannot find such a body of work in Perl.
I half-expected to find a module on CPAN, but alas...

If anyone if familiar with such a project, I would appreciate a
reference to it.

Just my 2-cents: Mostly Perl isn't used a lot for numerical
stuff. It's not because Perl isn't useful for such, it's just
that Perl's forte is text string manipulation. Each language has
its place -- Perl is great for string manipulation and general
"gluing" tasks, whereas FORTRAN, for example, is great for
numerical computation (FORTRAN will be many times faster than
Perl at heavy-duty numerical work). So one will be unlikely to
find a lot of numerical codes for Perl. If you would like an
optimizer, I have a Nelder-Meade optimizer implemented in Perl.
It's a bit long, so if you want it, email me via the link below.
....

 

[usenet]

Just my 2-cents: Mostly Perl isn't used a lot for numerical
stuff. It's not because Perl isn't useful for such, it's just
that Perl's forte is text string manipulation.

Understood and agreed. I was exploring the mathematical and mechanical
properties of the seige engine as explained in this _excellent_ paper:
http://www.algobeautytreb.com/trebmath35.pdf

The author uses Mathematica to do the calculations. I think Mathematica
is probably the best choice, but it doesn't work well for me because
it's a commercial license, it's (rather) expensive, and I don't know
how to use it. I'd much rather use Perl, and I'm not really concerned
about performance (it's just a hobby interest and an educational
exercise).

It does appear that CPAN offers computational modules to do the "heavy
lifting" of the required differential equations (Math::RungeKutta, for
example). And the bulk of the computations involves nothing higher than
basic trig. So I think the numerical tools are available in Perl, but
stringing them together is (for me) a bit intimidating for an all-out,
hinged counterweight, constrained & sliding sling trebuchet with a
cherry on top.

But I doubt I'm the first guy to think about trebuchet design in Perl.
I just wonder if anyone else has done more than just _think_ about it

 

[axel]

properties of the seige engine as explained in this _excellent_ paper:
http://www.algobeautytreb.com/trebmath35.pdf

is probably the best choice, but it doesn't work well for me because
it's a commercial license, it's (rather) expensive, and I don't know
how to use it. I'd much rather use Perl, and I'm not really concerned
about performance (it's just a hobby interest and an educational
exercise).

It does appear that CPAN offers computational modules to do the "heavy
lifting" of the required differential equations (Math::RungeKutta, for
example). And the bulk of the computations involves nothing higher than
basic trig. So I think the numerical tools are available in Perl, but
stringing them together is (for me) a bit intimidating for an all-out,
hinged counterweight, constrained & sliding sling trebuchet with a
cherry on top.

But I doubt I'm the first guy to think about trebuchet design in Perl.
I just wonder if anyone else has done more than just _think_ about it

Just make sure you get the calculations correct!

Inquest told of death by catapult

Story from BBC http://news.bbc.co.uk/go/pr/fr/-/1/hi/england/oxfordshire/4393524.stm
Published: 2005/10/31 15:30:20 GMT


A student died after being hurled by a catapult in an extreme sports
stunt, an inquest has been told.

Oxford University student Kostydin Yankov, 19, suffered multiple
injuries when he missed a net after being flung from a medieval-style
catapult.

Mr Yankov was part of the university extreme sports club - the
Oxford Stunt Factory - which was visiting the site in Middlemoor
Water Park in Somerset.

The inquest heard several members were concerned about the catapult's
safety.

Mr Yankov, from Bulgaria, known as Dino, was one of five members
of the Stunt Factory to use the trebuchet catapult on 24 November
2002.

Oliver Nelkin, who was due to jump after Mr Yankov said he was
alarmed that jumpers were landing at the very front of the net
rather than in the middle as intended.

Describing Mr Yankov's jump he said: "At some stage I saw Dino as
a ball in the air.

"He then missed the safety net but I couldn't say by how much. As
he hit the ground I heard a thud and then a second thud."

Paramedics rushed to the scene and Mr Yankov was taken to Bristol's
Frenchay Hospital where he later died.

The designers of the catapult - Richard Wicks, 33, from Burnham-on-Sea,
Somerset, and David Aitkenhead, 46, of Fiddington near Bridgwater
- were acquitted of Mr Yankov's manslaughter after a trial collapsed
at Bristol Crown Court last year.

The inquest, which is expected to last for three days, was adjourned
until Tuesday.

 

[usenet]

Just make sure you get the calculations correct!

It's possible to ride one of these things:
http://www.trebuchet.com/articles/ron/trebuchet.html

But it's one of the dumbest things I can think of. I plan to throw
16-lb bowling balls

Actually, FWIW (and I know this is getting a bit OT for Perl), I'm
interested in the effects of using a more flexible beam - I want to try
using 1" diameter oriented-strand fiberglass rod for the long arm. But
I need to figure the arm accleration to calculate an optimal release
point (when the rod is snapping back straight as the counterweight
starts to back-swing and the sling begins to slide off the release
finger). I also have been thinking about adding steel springs to the
sling tieback to get a little extra "snapping" effect at release. So
I've got a couple of extra parameters to consider which are not part of
traditional treb simulators (WinTreb, a-Treb, etc). A lot of this will
be trial-and-error in the field (optimal release angle, etc), but I
need to get "pretty close" before I begin, so I need to do bunch of
number crunching. And, of course, I prefer Perl because that's what I
have and what I know.

 

[Bob Walton]

It's possible to ride one of these things:
http://www.trebuchet.com/articles/ron/trebuchet.html

But it's one of the dumbest things I can think of. I plan to throw
16-lb bowling balls

Actually, FWIW (and I know this is getting a bit OT for Perl), I'm
interested in the effects of using a more flexible beam - I want to try
using 1" diameter oriented-strand fiberglass rod for the long arm. But
I need to figure the arm accleration to calculate an optimal release
point (when the rod is snapping back straight as the counterweight
starts to back-swing and the sling begins to slide off the release
finger). I also have been thinking about adding steel springs to the
sling tieback to get a little extra "snapping" effect at release. So
I've got a couple of extra parameters to consider which are not part of
traditional treb simulators (WinTreb, a-Treb, etc). A lot of this will
be trial-and-error in the field (optimal release angle, etc), but I
need to get "pretty close" before I begin, so I need to do bunch of
number crunching. And, of course, I prefer Perl because that's what I
have and what I know.

Actually, come to think of it, I do also have a translator for an
ACSL-like simulation language that would be good choice to
simulate a trebouchet. The translator is written in Perl, but
the output of the translator is FORTRAN-90 source code, which
must then be compiled with a FORTRAN-90 compiler to generate an
executable. Two versions exist: a standalone version that
creates plottable output as a flat text file of numbers, and a
Matlab version that interfaces with Matlab as a DLL (or "mex"
file) to form a rather nice setup in which Matlab does the
plotting and analysis. The translator works well, but is poorly
written (being among my first attempts at Perl years ago). It is
quite lengthy (over 100K of Perl code), so if you have interest,
email me at the link below.