PEP0238 lament AND Re: Case (In)sensitivity

Stephen Horne steve at
Mon Jul 23 23:33:56 EDT 2001

On Mon, 23 Jul 2001 18:35:35 -0500, "Chris Gonnerman"
<chris.gonnerman at> wrote:

>Both of these changes fall in the same group IMHO... they
>break a lot of code.

Case insenstitivity is also a serious issue for me - it might even
bite me for adopting a case-based naming convention as an apparently
sensible way to avoid future identifier conflicts - but I'm overloaded
by the division thing ATM. My brain has no room for further panic.

Anyway, it's probably less an issue for me than division - I use
integer division a lot, whereas defining different identifiers with
only case-changes to distinguish them is a definite no-no.

Case insensitivity on data (a subject line I glanced at briefly
mentioned insensitive dictionaries) would be a bigger issue - but
(while I'm too scared to check at the moment) I don't think that's the
point. The only nervous twitch is the thought that identifiers are
stored in dictionaries.

The argument for case-insensitivity (or - in my mind better still - a
case-consistency rule that rejects identifiers differing in case only)
is a very strong one. The in-twenty-years (or even 1 or 2 years)
argument has a lot more weight. The past experience of the industry
has shown that case insensitivity for identifiers and keywords
generally works - for many people, a major Unix/Linux wart is that
fact that the same command spelling - with a case change - can run a
completely different command.

Maybe it's just me not looking for the counter-arguments, but this
doesn't sound the killer that changing division will be for me. One or
two apps might fail from keyword clashes, but they shouldn't quietly
give out wrong results for anyone who names things sensibly. An
occasional bug report is a fact of daily life - it's the avalanche
that worries me.

There's still an argument for keeping backward compatability, but it's
both obvious and (personally) not that important - certainly it's not
going to give me a heart attack. I expect a lot of people will be
grateful to hear that that's all I've got to say on the topic ;-)

More information about the Python-list mailing list