[Python-Dev] Dafanging the find() gotcha

Tim Peters tim.one@comcast.net
Mon, 05 Aug 2002 19:17:04 -0400

[Andrew Koenig]
> Unless you're trying to teach the language to someone else, in which
> case you have to explain the behavior regardless of whether you've
> written programs that depend on it.

I would tell them that searching for an empty string is silly, that they're
never going to need to do it, but if they do then they should consider
whatever happens an accident.

> I doubt you've ever written a program that searches for the string
> 'asoufnyqcynreqywrycq98746qwh', yet I imagine that you would still
> object to a search function that throws an exception when presented
> with that particular string.

Of course, but I can easily conceive of *wanting* to search for
'asoufnyqcynreqywrycq98746qwh'.  Indeed, I just did a grep over my Python
code to be sure that I never had searched for it before <wink>.  But I can't
conceive of wanting to search for an empty string, despite effort after
suspension of disbelief.

> I'm not trying to be flip here -- I'm trying to make the point that in
> my opinion, having a uniform rule is preferable to catching particular
> cases that are sometimes mistakes.

The distinction between empty and non-empty is the only one being made here,
and (unlike picking on 'asouf'etc) is a natural distinction in its domain.