Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: st: Adjusting -tabout- to include top.tex and bot.tex lines automatically?


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   RE: st: Adjusting -tabout- to include top.tex and bot.tex lines automatically?
Date   Wed, 18 Nov 2009 13:01:21 -0000

I don't think that what Roy appears to be referring to is, for the most
part, mysterious or mystifying at all. 

I made a loosely similar point yesterday in pointing out that a
longstanding program of mine almost solved a recently posed problem (and
with one tweak, did so), so why were people not working from that? But
it's understandable, nevertheless. 

Inefficiency through duplication of effort is, in abstraction, naturally
to be regretted. 

But with the best will in the world, 

0. The same code segments -- or precisely what appear to be similar code
segments -- are very likely to be produced by programmers using the same
language. As Feynman said somewhere, the same problems have the same
solutions. I've been credited with tricks that I learned from Bill Gould
a decade ago, but only if he savages my code does he now get an explicit
Acknowledgment.  

1. User-programmers can easily not be aware of everything else that has
been done. I've used Stata pretty well every working day since 1991 but
I forget even things I've used a lot, and I imagine that many other
people would make similar claims.

2. Users really don't, and shouldn't, want to have to skip between
several programs if one program exists that is for them the union of all
their desires. 

3. Equally, it's not extraordinary that programmers want to write such a
program too in any given area. The implication of Roy's argument is that
a programmer should back off in so far as something already exists in a
given area. Excuse me, but no. For example, I like -dotplot- but it has
to me serious limitations, so that I wrote my -stripplot-, which
although not originally intended as such is now pretty much a superset
of -dotplot-. I would not accept a constraint that my program be
restricted to what -dotplot- cannot do. That way, we just end with lots
of little bits. 

4. In a very strong sense, we understand what we create (another
Feynmanism, somewhere). One of the reasons people write programs is so
they can understand exactly what is done. That's an intensely personal
process. Of course, there are always limits to that. I am very happy
that other people have written programs that I am content to use as
black boxes, indeed almost all of them. 

Nick 
n.j.cox@durham.ac.uk 

Roy Wada

Personally I am very mystified by a few people with a history
duplicating the same functionality found in other user-written
commands when in fact there are so many tasks and functionality that
remains unprogrammed and unavailable for a wider use. There's really
no need to beat the same topic over and over.


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index