RE: st: A comment on abbreviations

[Nick Cox <n.j.cox@stata.com>]

As maintainer of the FAQ, I find this a hard one to call.

First off, if Clarity and Obscurity are the candidates in an election, 
then everyone presumably votes for Clarity, so no surprise there.

However, we have not had advice about abbreviations in the FAQ, and I 
can't see that any advice we might give would in practice change 
anyone's habits, except that a few people might try to avoid 
abbreviations for a few days, but entrenched habits would then supervene.

Also, as maintainer, I have a personal bias in that I could not feel 
comfortable recommending something that I would have a hard time doing 
myself.

Svend's example is not very convincing: even if -prob- is not readable 
instantly as -probit-, then it takes only a little effort to find out 
what it does.

More generally, as has been pointed out, it is something of a service to 
people to remind them subliminally of possible abbreviations. You are 
going to have a hard time reading anybody else's Stata code if you don't 
quickly learn that many Stata people use abbreviations like -d-, -l- and 
-gen- all the time.

So far as the FAQ is concerned, I can't see a clear case for change. If 
you agree with Svend, you can of course use full names. If you 
abbreviate, we are not going to say that is wrong.

Naturally, this question intersects with other enjoyable little debates, 
such as whether abbreviations were a bad idea on balance, and why do 
most Stata users seem to abbreviate -d- and -l- harshly, -gen- less 
harshly and why do most Stata programmers seem to write out -local- in 
full, but let those be.

Nick
permanent address:n.j.cox@durham.ac.uk

Svend Juul

Abbreviations make it easier to write a text, but more difficult to read 
it. Especially if your experience with the language is limited, 
abbreviations make life difficult. I have the impression that the use of 
abbreviations at Statalist have increased during the last months, and 
recently a response included this:
   prob f p w

It means
   probit foreign price weight
- and that is much easier to read.

The Statalist FAQ includes good advice, mostly to less experienced users 
seeking help. I think it should also include this advice to the more 
experienced users offering help:

   Be helpful. Avoid abbreviations.



*
*   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/