st: Re: mata code

[Kit Baum <baum@bc.edu>]

Don't hold your breath. At the London SUG meetings, StataCorp speakers 
presenting mata told us that they considered mata code for "official" 
Stata routines to be in the same category as the C code for "official" 
Stata built-in routines (e.g., the code underlying -regress-). Thus, 
they will distribute it in compiled form (in mata libraries), but not 
expose the source code.

We might hope that the Stata Journal's regular "from the help desk" 
feature will include some examples of how mata can be used to solve 
econometric programming problems. That feature often is accompanied by 
ado-file code, and one would hope that such articles -- which do not 
discuss official Stata features -- might also present the mata source 
code for the edification of user/programmers. It is a great advantage 
of Stata that much of the code is exposed, and the best practices 
embodied therein can be studied by user/programmers. It would be indeed 
unfortunate if a larger and larger fraction of Stata's commands are 
made inscrutable by means of mata (while giving efficiency and speed, 
removing the models by which many of us have written add-ons to do what 
official Stata does not).

Kit Baum, Boston College Economics
http://ideas.repec.org/e/pba1.html

On Jul 7, 2005, at 2:33 AM, Rob wrote:


> Also, in general, I'd like it if StataCorp would distribute both the
> compiled libraries and the source code for routines such as these.
> One of the nice things about implementing commands as .ado files is
> that users can look at the code and see exactly how the commands
> work.  Mata seems like it'll be a great tool, but it'll be a lot more
> useful if users can see the source code than if we only get to see
> compiled libraries.
*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/