Why python???

[Michael Peuser]

Hi Moshe,
though I decided to no longer answer to that thread, I will make an
exception ;-) Can you possible explain why my benevolent and pro-Python
attitude (though not without critical touch) can be so mis-interpreted??

----- Original Message -----
From: "Moshe Zadka" <m at moshez.org>
To: "Michael Peuser" <mpeuser at web.de>
Cc: <python-list at python.org>
Sent: Saturday, September 06, 2003 7:29 PM
Subject: Re: Why python???


> On Sat, 6 Sep 2003, "Michael Peuser" <mpeuser at web.de> wrote:

> > ...One of the things that impressed me in Python was that it is
> > aware of complex numbers and it has NumPy (which in fact is *not*
Python).
> > Though I think that this very rarely is stated as one of Python's
> > overwhelming assets and most people even wonder what this can be used
for

> You're obviously wrong on that count.

Well, how can you say that?? May be I err, but because you and I and a *lot*
af people use NumPy, that does not prove that - to my experience - *most* do
not use it. Period.

> Or, rather, it might be that this
> is rarely stated as an advantage -- but this is because it is irrelevant.
> I do know, for a fact, I do recommend it as an advantage when people are
> asking me what to use. For example, a friend of mine wants to do 3D
> sketches for some notes about elliptic curves he's preparing, and one
> of the advantages I stressed when recommending Python was that it is
> easy to do stuff like "take this array of vectors and give me an
> array of vectors multiplied by this matrix". Of course, I realize that
> most people think that multiplying vectors by matrices is stuff nobody
> needs, and I would *not* stress it as an advantage if I was trying
> to explain to my graphic designer friend why Python is a good language.
>
> As to NumPy being "not Python", I hardly see how this is relevant.

No it is not very relevant! It is just the observation that you could more
or less easily include it in every language with some heap-support. So I
would not overstress it as such, as I would not stress the fact that Tcl/Tk
is tightly coupled to Python - which is not considered an advantage by
everybody.

> NumPy is enjoying *excellent* syntactical support from Python [for
> example, a[3:5,...,4:6] is actually working code which does what
> it looks like, namely gives a multi-dimensional array with the same
> dimensions, except the 0th dimension only gets the 3-4 layers and
> the last dimension gets the 4:6 layers (try to do this in other
> non-numeric-specific languages!)] and is fairly easy to install.

This is *just nice*

> You may not be aware of it, but it is eventually slated to get in
> core Python -- it's just that the NumPy guys themselves wanted a do-over
> [in the form of numarray] before it is blessed into the standard
> distribution. But even until then, it is easy enough to install
> on most operating systems, so I consider that as a core advantage
> of Python itself.

Though I am looking to numarray from time to time I was not aware of this;
it is certainly a good thing. Beeing included in the standard distribution
is a kind of promotion, but does not make it "Python", as the STL is not
C++.

> > However the *real* programs then are the mathematical packages, OpenGL,
> > simulation and visualisation.
>
> That's a lot like saying that the "real" programs are the CPU designs.
> I am betting fairly certainly that Intel has spent more money on its
> CPUs's design than, say, the OpenGL team has on its software -- and
> this gets even more ridiculous if you start considering how much money
> nVidia spends. What it comes down to is that I don't care if you call
> what I do "real" programming or not -- I care only for getting a job
> done. If I write a statistical analysis program which is much shorter
> than the NumPy code, does it make my software non-real? I don't know.
> But when I did write statistics code, I was sure happy I could play
> with parameters and different filtering algorithms faster than I could
> do it in any other way [including dedicated statistics software --
> we had no SAS and SPSS is a piece of crap].

Maybe I should explain the background of my arguing which I had not done in
the NG thread: We have full-time programmers and domain-specialists who are
(more or less) forced to do some programming because they couldn't work
otherwise. There are arguments whether it is necessary or useful for both to
use the same programming language since we had programming languages. This
is how BASIC came into existance which no *real programmer* would ever touch
(as programmer that is). From yout point of view as "scripting user" you are
absolutly right. From the point of view of a "tool making" programmer there
are more complex considerations. They rollup to the point you already
stated: "I care only for getting a job don." But the programmer has a
different job...

Kindly
Michael P


>
> --
> Moshe Zadka -- http://moshez.org/
> Buffy: I don't like you hanging out with someone that... short.
> Riley: Yeah, a lot of young people nowadays are experimenting with
shortness.
> Agile Programming Language -- http://www.python.org/