[Edu-sig] Calculus with Stickworks (more Gnu Math)

kirby urner kirby.urner at gmail.com
Sat Sep 30 20:10:51 CEST 2006


On 9/30/06, kirby urner <kirby.urner at gmail.com> wrote:

> if it could be taught more like a computer science.  That's what I'm
> already up to, plus having the most active Bucky site on the net (BFI
> didn't exist, no Bob Gray's site -- Chris Fearnley with his FAQ was
> one of the first).  Only then do I come across Python.
>

BFI didn't yet exist as a hosted domain on the W3 I should say.
Kiyoshi and his Philly friends got the bfi.org moniker secured (John
Ferry catching from inside the BFI, then in LA), while I webmastered
the first primitive website (still retrievable via archive.org -- note
they've so-far kept with the tabs format, plus built something
agreeably interactive (the web is less about push, more about pull
these days, which I like -- more people publishing their own content,
not just passively browsing (of course one wants both active and
passive modes)).

You'll find me in an old FoxPro Advisor, yakking up Chakovians
Coordinates (calling 'em Quadrays), long before I'd stumbled across
Python.  I'm trying to think how Microsoft could make VFP affordable
to schools, for the kind of math-through-programming exercises I'm
thinking kids'd love (no, not every kid, and no shame in preferring
football).  But VFP is already a twisted story by then, having been
appropriated by MSFT as a "dBase killer" (Borland went with
Ashton-Tate's branding), touting Xbase as one of the Pillars they'd be
supporting (right up there with VB and C/C++ I think it was).

Obviously BS in retrospect.  XBase is not general purpose like VB and
not low level like C/C++.  It was an early agile, interactive from the
beginning, and when upgraded to an OO, was actually more sophisticated
than VB -- we *defined* our own classes in a Visual Studio like IDE,
didn't just use blackbox DLL objects supplied to us by COM or whatever
(though we could use those too, and still do).  VFP has always been
better than average, by Microsoft's own standards (no shame in having
an "average" that you're sometimes "better than" -- I was not saying
this to be disparaging of MSFT).

Anyway, yes, Chakovian Coordinates, centered at (0,0,0,0).  Tom Ace
chimed in with some 4x4 rotation matrices and wow, we had like this
whole little vector algebra thing going on Synergetics-L.  Someday,
we'd capitalize on that even further.

The Gerald de Jong entered my life, all hyped about Java.  He flew out
to JavaOne.  The guy was already a Canada-trained mathematician and
C/C++ guru.  It's not like he was just cutting teeth.  He had this
vision of something called Elastic Interval Geometry and counted both
Kenneth Snelson and Bucky Fuller as inspirational in that regard.

Following Gerald's lead, I saw the wisdom in using a free language,
especially one that'd work over the web, so dove into Java pretty
seriously.  I forget exactly how that connected me to Python, but it
did.  Linux was hardly usable as a GUI desktop yet, so all this is in
Windows, dating back to early DOS, and in my case, yes, even CP/M on a
Zorba.

My trajectory from FoxPro through Java to Python is well chronicled in
a multipart Geometry through Programming paper I published to
4dsolutions.net.  I'm always focused on the same Polyhedra, using
Chakovians or whatever, but I keep doing it over in a different
language.

Towards the end, the Scheme Team was generously coaching me on how I
might continue my arc (these were the early edu-sig days, plus I was
already posting to that same Math Forum forum), and implement the core
ideas yet again in that language.

I'm not saying it's a bad idea.  However, for all my digging into PLT
Scheme, I think by this time I was too mired in other paradigms.
Lambda Calculus was never my focus, even at Princeton, where a lot of
the original work was done (as I learned only later).

Here's that multi-part geometry paper (VFP -> Java -> Python arc i.e.
not really the Shuttleworth pipeline, especially when you count the
early APL and the ongoing J branch, plus I've dabbled in Perl, C and
Pascal, plus did FORTRAN for money (the only way I would)):

http://www.4dsolutions.net/ocn/oop.html

Kirby


More information about the Edu-sig mailing list