[AstroPy] Summary of Python related events at ADASS

[Rick White]

>From owner-astropy at stsci.edu Tue Nov  2 10:36 EST 1999
>From: Paul Barrett <Barrett at stsci.edu>
>To: astropy at stsci.edu
>Subject: [AstroPy] Summary of Python related events at ADASS
>
> [...]
>
>As an aside, a similar utility for converting IDL programs to Python
>would be a nice asset to the AstroPy toolbox and provide valuable
>experience for future enhancements to the Numerics module,
>particularly for image processing.  Anyone interested in taking on
>this project?

I've been thinking about doing this myself.  I think IDL has a much
simpler syntax than the IRAF CL, so the parsing would be easier.  The
Python code generation would currently be hard, mainly because of
missing features in Numeric like the ability to index arrays with
arrays and the "inverse of take" functionality.  That would require the
generation of a bunch of messy code for simple IDL statements, which
would probably be off-putting for IDL users.

Once the few holes are filled in Numeric, I think it would be possible
to produce clean, readable Python translations of IDL procedures.  Then
the effort would focus on creating a useful subset of the IDL built-in
library routines, which could be shared among many developers.

It took me a few weeks to implement the CL-to-Python translator using
Aycock's SPARK framework, and I suspect doing the IDL version would be
a bit quicker if the Numeric facilities were available...

- Rick

---
Richard L. White    rlw at stsci.edu    http://sundog.stsci.edu/rick/
Space Telescope Science Institute
Baltimore, MD
_____________________________________________________
AstroPy mailing list  -  astropy at stsci.edu
http://lheawww.gsfc.nasa.gov/~bridgman/AstroPy/