Linux Journal writeup



On Fri, 2011-07-15 at 07:07 -0700, Richard Fateman wrote:

> Both, I think.  The idea that Python is suitable for a front end to 
> computer algebra is hard to entirely endorse,
> considering that after boasting about it, they then find it unsuitable 
> themselves without a preprocessor
> to handle (as I recall) the fact that integers and arbitrary-precision 
> integers are different types. Maybe
> some other things.

As a recent "refugee" from Sage, I far prefer the Maxima language for a
front end to Python. Don't get me wrong, Python is a great language for
some applications, this just doesn't happen to be one of them!

> 
> >   Typically math systems
> > (symbolic and numerical alike) have a user language which is
> > just cobbled together as an afterthought. Python has, at least,
> > the advantage of being more nearly complete and more
> > carefully designed.
> I think that evaluating Python as a language to convey essentially all 
> of mathematics, Python
> is going to appear cobbled together.  Also note that Python has changed 
> from time to time in
> fairly significant ways.
> 

It has also changed relatively quickly and different platforms don't
always keep up with each other; different Linux distributions often have
a slightly different version as "standard".  Changes can cause a lot of
hassle without necessarily being of any benefit to a project.  CL has
many dialects but I don't think the language is going to change much!

> One claim (maybe marketing hype too? I don't know first hand) is that 
> there is ONE implementation
> standard for Python on all platforms, and so there is no need to deal 
> with GCL CCL ECL Scieneer, Allegro...
> dialects.
> 

I would describe this more of a pious hope than a claim!

Cheers,

Bernard Hurley.