Devon McCormick <[email protected]> writes:

> Bill - my suggestion is that you find out what they want to do and help them
> tackle real problems.  To this end, it's important to make them aware of the
> various data I/O facilities - not just Tara, but also 'csv' and ODBC
> access.  Also, the 'images' library is useful.  In a related vein, I found
> Cliff Reiter's work on filling in missing image data - see
> http://ww2.lafayette.edu/~reiterc/mvq/bafrmid/index.html - very impressive.

Devon,

Yes, I like Cliff's image data example.    

I have asked around as well as listened to what they've told me.  I
think they first want a general (and brief -- like J) introduction to
what this is, so I'll do a few minutes that culminates in a few very
brief stats examples from Keith Smillie and John Baker.  Then I'll
probably send them off to do chapter 1 of Henry Rich's introductory lab
(we only have 30 minutes total a week, so some of that will turn into
"homework," if they're interested).

I think scripts and simple plotting may come in the second week, and
accessing real data may come in the third.  I'm still listening, though.
I don't think they're interested in number theoretic stuff and the like.

> A recurring problem we talk about in NYCJUG is finding what's available -
> there's a wealth of material but it sometimes is hidden because of the way
> we talk about it - we have our own peculiar math-centric vocabulary.

There is some great older material, too, but it takes a bit of
experience to figure out changes in the language since the last J7 (I
happen to have some of the old spiral-bound books still).  There aren't
many, but the change in the rank of @ and a few things to do with
conjunctions come to mind, IIRC.  (For newcomers, you'll probably have
to work pretty hard to find old material that has these problems, so
don't worry about it.)

> I'm intrigued that there's a couple of programs - mentioned on the J Wiki -
> written by people we've not heard from before or since.  Obviously they were
> able to accomplish something with J even without knowing the language well.
> I'm thinking in particular of this implementation of diffusion-limited
> aggregation - http://www.jsoftware.com/jwiki/Studio/Gallery/DLA - which I've
> used as a starting point to try to demonstrate how code can be made more
> J-like to take advantage of its strengths - and this implementation of a 2-D
> convex hull: http://www.jsoftware.com/jwiki/DevonMcCormick/convexHull .

Thanks for the good suggestions, and thanks for the link to the convex
hull piece.

Bill
-- 
Bill Harris                      http://facilitatedsystems.com/weblog/
Facilitated Systems                              Everett, WA 98208 USA
http://facilitatedsystems.com/                  phone: +1 425 374-1845
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to