[AstroPy] AstroLib vs Astlib vs APLpy vs ...

Perry Greenfield perry@stsci....
Thu Jun 25 07:40:31 CDT 2009


On Jun 24, 2009, at 7:53 PM, Marshall Perrin wrote:

> Hi all,
>
> I'm a member of that great teeming horde of not-really-using-Python- 
> yet astronomers, slowly trying to wean myself from IDL and onto  
> Python.  I've written a handful of little things in Python but want  
> to dive in fully for my next project. As part of that process, I'm  
> trying to figure out what the real must-have list of python modules  
> is for everyday astronomical tasks.  In IDL, there's sort of a  
> consensus library that most folks use, at least in part: the Goddard  
> IDL Astro library, Craig Markwardt's MPFIT library, Dave Fanning's  
> display routines, and so on.
>
> It doesn't seem like Python yet has that consensus, once you get  
> much beyond the triumvirate of scipy, numpy, & matplotlib.  Am I  
> wrong? As far as I can tell there are several different libraries  
> aspiring to be "*the*" astronomy library in Python - AstroLib and  
> AstLib both, for instance.  You can display nice plots with  
> AstLib.ImagePlots or APLpy, manipulate coordinates with  
> AstLib.coords or PyWCS, and so on. Which of these should I prefer,  
> as a relative python neophyte? Which have the most developer  
> momentum behind them right now?  It's all a bit confusing, and I'd  
> rather try to spend my time being confused about my science objects  
> rather than my software choices. ;-)
>
> It seems like there was a brief attempt to start a discussion about  
> this problem here last year (see http://mail.scipy.org/pipermail/scipy-user/2008-April/016062.html 
>  and followups) but as far as I know, nothing really came of that.
>
> I don't expect any easy answers here. I'm just curious to hear what  
> 'the community' is mostly using these days - if there even is a  
> consensus! - and what the recommendations are for someone trying to  
> make the big switch away from IDL. Thanks very much!

You certainly point to an important issue and I won't say it isn't a  
problem at some level. On the other hand, I see it as a good sign in  
some respects. It wasn't all that long ago that no wcs library was  
available. And now there are too many :-) That's indicative of much  
greater activity in this area and I think that bodes well for the  
future if it means some confusion right now.

I was hoping soon to try and start some effort to try to integrate  
these efforts, if not into common packages, a common place that people  
could obtain them from. I think over time, these different approaches  
will eventually settle on one or some merging of the existing ones  
once there is a chance to see what the merits of the alternative  
approaches are. It's true that this means some annoyance for users in  
not knowing which one they should use for a while or the possibility  
that the final community package may require them to make changes to  
their scripts. Hopefully this situation won't last a very long time  
for wcs packages. But it isn't always a bad thing for two or three  
different approaches to be tried initially.

Perry





More information about the AstroPy mailing list