Am 29.11.2014 um 02:44 schrieb Richard Fateman:
On Thursday, November 27, 2014 10:35:34 PM UTC-8, Joachim Durchholz wrote:
Awesome.
The papers I've read have been almost exclusively from the theorem
proving
world.
I think you should be mostly fine working off these.
I disagree, unless you are able to find much better papers than I have seen.
Well, maybe not "just fine", but simply picking up Lisp code isn't going
to work *that* well (and would be quite pointless), so what alternatives
would be there?
Essentially it's all tree matching of some kind. Things will start to
diverge as soon as domain specifics start to matter; it would be nice to
have a not too domain-specific basic building block and add the various
strategies on top of that.
I think the "various strategies" are very important, and not just add-ons.
Of course they are important.
It would still be nice to be able to use different strategies, depending
on what the particular algorithm tries to do, and on experiences with
speed and maintainability of Python code.
There is a pattern matching program in Peter Norvig's book on Artificial
Intelligence Programming ; this book will also teach you Lisp.
I agree that learning Lisp is a good idea. I'd recommend complementing
that with Haskell (e.g. "The Haskell School of Expression", though it's
a bit superficial), and http://www.info.ucl.ac.be/~pvr/book.html to see
what else is out there.
However, that's about learning programming. Which is a good thing to do
in general, but it does not really help with designing Python classes
for SymPy.
For that, firstly, you'd need to understand what exactly each design
decision in Peter's (or anybody else's) code was intended to achieve.
For that, merely learning Lisp is not enough, you need to have mastered
it - which is a few years of experience away for most of us (except for
you, I assume, and maybe the odd one or two lurkers on this list).
Second, you need to understand all about how to design the equivalents
in Python, dropping those design patterns that don't work in Python, and
replacing them with something that works as well. (For example,
everything macro would have to be replaced or redone; is there anything
relevant left if you take out macros in the useful Lisp work?) So you
need to be a Python master as well. Given your record I doubt that you
are, I certainly am not, and those people around here that I'd consider
Python masters, I doubt that any of them is also a Lisp master.
Third, the resulting code needs to be maintainable by people who do not
know anything about Lisp, and barely enough of Python to become a GSoC
student.
I have serious doubts that Lisp style is going to work for that, at all,
even if you let a Lisp And Python Master do the translation.
So right now, my point of view is that trying to get that (doubtlessly
very valuable) Lisp experience into SymPy, while it could do all kinds
of good, is probably not practically possible.
Unless somebody can delineate a clear path to overcame the various
obstacles I just enumerated.
I highly recommend it. In fact, I would go further than that, and say
that I would expect you to re-invent much of what he writes out,
or alternatively, have a severely deficient program.
Sure, but if his book makes you learn Lisp, the presentation is heavily
biased towards Lisp, and much of what he writes won't be transferrable.
My recommendation would be to reinvent first (and discover how the
Python patterns for that particular field work), then read the Lisp
equivalents, then transfuse the ideas.
The point being that once enough experience with the Python patterns is
available, it's easier to see what to transfer and what to better leave
untransferred.
YMMV :-)
Regards,
Jo
--
You received this message because you are subscribed to the Google Groups
"sympy" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/sympy.
To view this discussion on the web visit
https://groups.google.com/d/msgid/sympy/547958A9.8090207%40durchholz.org.
For more options, visit https://groups.google.com/d/optout.
