Case-sensitivity: why -- or why not? (was Re: Damnation!)

[Tres Seaver]

In article <slrn8jrb37.b4q.wtanksle at dolphin.openprojects.net>,
William Tanksley <wtanksley at bigfoot.com> wrote:
>On 06 Jun 2000 17:34:20 -0400, Andrew M. Kuchling wrote:
>>wtanksle at dolphin.openprojects.net (William Tanksley) writes:
>>> I'm not the one that's claiming that removing case-sensitivity will ruin
>>> Python.  My claim is very simple: first, it's a minor change for us to
>>> make, and second, it helps newbies.
>
>>Like Aahz, I don't think the change is that minor.
>
>I suppose we'll just have to agree to disagree, then.  Because I can't see
>the change causing me to care -- my coding style will hardly change.

*Sigh* It's not *your* style I'm worried about -- I've seen stuff you
post, and *you* wrtie cleanly.  I have significant experience, however,
in working in *large* (100K+ LOC) bodies of code written in a case-
insensitive language (ObjectPascal) by case-negligent coders.  The
readability of case-mangled code, like that of inconsistently indented
code, is *horrible*.  "Coding standards" get ignored by this grade of
"programmer";  compiler errors don't.

>
>Now, the change WILL cause old code to not work -- but in case you missed
>it, it's been established that Py3K is going to do that.  Adding or
>removing this feature will not fix the code.
>
>>I'm also doubtful of the "helps newbies" claim; unless there have been
>>some other citations in this thread that I missed, we have exactly one
>>study using exactly one programming environment that reported CS
>>(case-sensitivity) was a problem, and even that report says "most [users]
>>continued to type case-incorrect tokens in their programs for a short
>>period", which implies that they did eventually stop.  The evidence is
>>hardly overwhelming, and basically purely anecdotal.
>
>Now here we can make some progress, I think.  Once CP4E is funded, I would
>hope that this is something they would do some work in.  Hopefully before
>the change is finalised in Py3K.
>
>However, from my experience teaching people programming and typing, people
>don't normally think or communicate in a case sensitive way, and a large
>portion of people have a really hard time doing so.  I've hypothesised
>that this is an artifact of visual versus auditory learning.

Readability is an important element of both learnability and usability
of a language;  case-insensitivity reduces readability in order to enhance
writability.  Ought we make it easier to write sloppy Python than
is presently possible?  I don't see how that can be a win, except for the
greenest of 4E's.

>>IMHO, if GvR was inventing a completely new language for teaching,
>>then it would be worthwhile to debate whether it should be CS or CI.
>>But he's not; Python is an implemented language with a sizable and
>>growing user base that's used to CS, and a number of published books
>>and other materials that will be invalidated by switching to CI. 
>
>And the fact is that they're going to be invalidated anyhow.  Nothing we
>do to CS will change that.  Part of the reason for Py3K is to allow Guido
>to make large, backwards-incompatible changes.

To what end?  I still have seen nothing except the anecdotal evidence from
the Alice project for the claim that CS reduces the overall usability, or
even learnability, of Python.

Gratuitous-incompatibility-buys-us-what?'ly

Tres.
-- 
===============================================================
Tres Seaver                                tseaver at digicool.com
Digital Creations     "Zope Dealers"       http://www.zope.org