On 11/27/02 15:28, Nick Cox wrote
>Richard Herrell
>>
>> On Wed, 27 Nov 2002, Nick Winter wrote:
>>
>> > > -----Original Message-----
[snip...]
>
>I am still not clear whether -x- and -y- are variables,
>and whether this all -- in Some Alternative
>Software -- is tacitly looping over
>observations. (In my case, please assume
>complete ignorance of SAS.)
>
>I will also add that I don't find often myself
>wanting to use completely different
>commands for different subsets of a variable.
>Alternatives could depend on what those
>commands are.
I've found one situation where this would be nice: when reading in really
nasty data files. Two places where this happened in the past to me were:
Some really badly planned web forms, where I had to translate the
resulting very messy, non-delimited emails. From what I heard (from a guy
who had worked at SAS), SAS would be good for this, because of the
ability to evaluate each incoming line and process it differently
depending on its contents. I tried fixing things using lisp, ended up
learning Perl and later wished that I had learned Python first.
Some badly made datasets from a weather data repository where it would
have been nice to bring in the data line by line.
Now, while I sometimes wish that Stata could do this type of
manipulation, I generally prefer to have Stata do what it does well
(statistical analysis), and use other tools for complicated problems
which Stata does not do well.
The same fellow who told me about how well SAS would have fit my needs
also said that inside the SAS corp, SAS had to be called SAS (rather than
its old name Statistical Analysis System), because the SAS folks regarded
it as a data management package far more than a statistical package.
As for speed - in the distant past (oh, say, back in the mid 90s), I
recall writing Stata code which would do in 15 hours what SAS would do in
4 days. Of course - I suspect that the SAS programmers whose code I was
duplication were not particularly attuned to being efficient. Things
could be very different now.
Bill
*
* 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/
