Re: st: RE: accessing a variable's value WRT its attached value label

[Phil Schumm <pschumm@uchicago.edu>]

On Mar 29, 2006, at 4:13 PM, Nick Cox wrote:
> The question in return has to be: readable by whom, or more  
> precisely by what class of program reader?  what can be assumed of  
> their Stata experience and competence?

Excellent question -- I should have been clearer.  I was actually  
referring both to myself and to a fairly large group (~6-12) of Stata  
users, ranging in experience from novice to fairly advanced (but none  
experienced in writing his or her own ado files or, perhaps more  
informatively, none who has picked up [P]).  But your point is well  
taken; what might be considered readable may well differ between  
different groups of individuals.

I guess what I was really trying to avoid was the inline expansion of  
an extended macro function; in my experience, many users find this a  
much bigger jump than just learning how to use a simple macro (e.g.,  
within a -foreach- loop).  This is especially true with the second  
colon and/or if you throw in additional levels of nesting, e.g.

foreach var of varlist var1 var2 ... {
    count if `var' == "yes":`:val l `var''
}



> You don't say anything about comments.  If inscrutable code is your  
> concern, tuning a comment to give an explanation at the right level  
> is the simplest thing.

Absolutely, though your phrases "tuning" and "at the right level" are  
pregnant with implications.  As you know, good commenting style is  
difficult for many programmers to learn, let alone people who code  
only very occasionally.  Moreover, too many comments (and/or comments  
which are excessively long or poorly crafted) can actually decrease  
overall readability, and in no case is commenting a substitute for  
writing tight, well organized code.  Nonetheless, part of my concern  
in this case was that the most inexperienced individuals in the group  
at least be able to look at the code and get some sense of what is  
going on, and a brief, well-written comment at or near the command(s)  
in question would certainly facilitate this.



> 1. As a Stata user who also writes programs, I find the use of a  
> local macro and of an extended macro function unthreatening, but I  
> still have to work a smidgen too much to realise that
>
> `: val l y'
>
> is a minimal abbreviation of
>
> `: value label y'
>
> Having also worked with highly concise languages, such as J, in  
> which almost all primitives are one or two characters long, I have  
> to say that readability does often hinge on avoiding minimal  
> abbreviations.

Thanks for this reminder.  Although I'm not among those Stata users  
who write -g- for -generate- (personally I prefer -gen-), I must  
confess that in some cases I've gotten in the habit of using minimal  
abbreviations in other contexts without really thinking, partly, I  
guess, based on a rather unconsidered belief that a shorter line was  
in some sense "cleaner" (i.e., less "ink") and therefore better.   
However, I agree that the second option above is certainly more  
readable than the first, and am inclined to try and retrain myself  
away from my current behavior (at least in some cases).



> 2. The construct
>
> "label":valuelabelname
>
> is prominently documented, but my impression is that it is fairly  
> rarely used.

I agree, certainly if not among active members of Statalist, then  
among the majority of "regular" users.  Moreover, unlike inline  
expansion of extended macro functions, I would argue that this  
construct should be learned (and used) by everyone because of its  
potential to help avoid simple mistakes (e.g., forgetting whether 0  
is no and 1 is yes, or 1 is yes and 2 is no, or wait, wasn't that...).

As always, thanks for your words of wisdom.


-- Phil

*
*   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/