You are correct that -lookfor2- is not registered.
What you do you infer from that? You can certainly
infer that the author (myself) did not register it!
I have the impression that -cmdname- was largely a good
idea in principle, and a near total failure in practice.
I could be wrong, but it seems many user-programmers
either did not notice -cmdname-, or did not adopt the
practice of registration.
I am as much to blame as anyone, as it never became
part of my standard routine.
Also, I have not heard Stata Corp people mention
this either in public or in private for some
years. Perhaps they have forgotten too, or
accepted that the experiment just didn't work.
Not a big deal.
One principle announced
was that "The registry represents a loose agreement
among Stata programmers to try to name new commands
uniquely", but that was announced from on high,
and perhaps wasn't the right psychology, not that
it matters directly.
Actually, I think there really is a loose agreement to
this effect; it is just that the idea of needing
Stata Corp approval for your precise names was perhaps a
little too much to swallow.
In practice, clashes between program names do
arise occasionally and are usually easy to resolve;
in particular, if necessary, user command names _must_
yield to official command names. Make no mistake:
the latter must be a golden rule. But even
registration did not rule out the possibility
that a name accepted by Stata Corp would not
be grabbed back by them at some later stage.
And again, Stata Corp were right to claim
that, and to emphasise it, but it lessened the
appeal of registration.
Looking at the rules announced with -cmdname-,
I would say that they all have considerable
justification, but I have felt free to ignore
them on occasion, and others have too.
In what follows, _me_ means _you_ as well:
1. You may not register a name already registered.
This would seem binding, except that a name
used for some user program now moribund, or more generally
for a user program not used by _me_, is of no concern
2. You may not register a name that appears in the English-language
This has often been ignored, at least by
myself. After all, the English language
belongs to _me_ as much as to Stata Corp. (More so!)
Sometimes it seems that a word really is right for a program,
and -- especially if you think it's quite a nice
little program -- an ugly name seems wrong. Sometimes
you then have to give in, as Stata Corp later use the word
themselves, but not always. After all, there are lots
of good shortish words in the English language.
(Of course, Stata program writers can also use names in
Spanish or Elvish or whatever else they want. An interesting
middle case is _Greek_ words, which so often have
3. You may not register a name that is less than 4 characters long.
This has also been ignored, especially
privately for convenience abbreviations. Of
course, private programs are of no direct concern here,
except insofar as they sometimes are mentioned
as answers to questions on Statalist. (Nick Winter
has a program called -c-.)
4. You may not register a name that is a word of
statistical jargon that is in common use.
Similar comments to 2.
I am not trying to justify selfishness or petulance,
just trying to explain, anthropologically as it were,
why the tribe never really adopted this practice,
so far as I can tell.
There is also a much better answer: -findit- tells
you directly whether a program name is in public use.
That cuts through the whole question of names registered
or not, approved or not, or whatever. I suggest that
easy search of the internet finally made -cmdname-
> Thanks Nick for naming the command -lookforit-.
> I'll will document it and submit it to the SSC archive.
> I just discovered -cmdname- written by Stata Corp back in 1999
> that can be downloaded:
> . findit cmdname
> This command allows you to check to see if a command name
> is already registered at Stata and if not allows you to register
> it in your name with your email address.
> I noticed that -lookfor2- is not registered.
* For searches and help try: