[SnapPea-planning] priorities

[Nathan Dunfield]

Bill wrote:
> So --- it's not obvious to me the extent to which it's better (in the
> long run) to separate the different aspects of geometry and topology,
> communicating perhaps by  human-legible and editable ascii files vs.
> the extent to which it's better to aim for a single integrated master
> program whose inner workings are more shielded.

I think that there's a number of forces that cause mathematicians  
collectively to produce a lot of small, specific programs instead of  
integrated work.   For instance, mathematicians typically have  
limited programming skills, and so can only work in one or two  
languages that might well not be that of the more integrated program  
in the area.  Also, If you only need some small thing, learning the  
basics of developing for an integrated system can be a lot of overhead.

The good news is that, and this is the lesson of Sage, many small  
programs can be integrated into a coherent, useable larger program.    
I've done a lot of computations of this mix-and-match type recently  
(e.g. my paper with Dinakar), and Python is very good at allowing you  
stitch things together.

	Best,

	Nathan

P.S. to Nathaniel or Dylan or whomever is maintaining this list:

Could messages sent to the list have the "Reply-To" field set to  
"snappea-planning" rather than the author of the message.

In sending this one, I almost forgot that I needed to reply to all so  
it would go everyone, not just Bill.