*blink* That's neat.

[Martijn Faassen]

Fredrik Lundh <effbot at telia.com> wrote:
> Martijn Faassen wrote:
>> > I'm not sure Alice has much of a "result" to offer on the subject. As
>> > far as I can tell from the paper I have read, doing away with case
>> > sensitivity was just the nearest hack available, not a decision based on
>> > scientific analysis of the whole interface problem.
>> 
>> True; I don't think there was done any double-blind research into this
>> phenomenon.

> I'm probably just dense, but how would you design such a
> double-blind test?

Construct two Python interpreters (or environments). One is case-sensitive,
the other is case-insensitive. Then randomly distribute them to your
newbie programmers (you need a statistically significant batch of these;
don't know how many that would be). Set them a number of programming
tasks. See how quickly and how well they complete the tasks, and perhaps
do some measurement of their 'frustration level'.

Perhaps to make it even more confusing, instruct half the amount of
programmers about case-sensitivity issues (even those that end up
using the case insensitive interpreters; you don't know anyway).

That would seem to be a fairly double-blind test to me, right? :)
Also fairly tough to be.

>> It was just an anecdotal observation, perhaps in a more controlled setting
>> than ours, but probably still doesn't weigh much more strongly than our own
>> anecdotal observations.

> As far as I understand, the Alice researchers observed that
> errors caused by case *mismatches* are a real and important
> problem.  They don't claim that there are no better *technical*
> solutions than the one they implemented.

Yes, but those observations were probably anecdotal, right? They'll have
more experience with this than we do, and they'll have dealt with more
newbie programmers. But they'll have a bias just like anyone else. So
their anecdotal evidence, while stronger than our various claims, does
remain anecdotal. In science this means you can be inspired to form a
theory on it, but it doesn't automatically means your theory is closer
to the truth; you need empirical testing to strengthen it. (but *having* a 
theory is probably at least better than having no theory at all)

> (iirc, alice just casefolds everything in the "reader".  Which
> reminds me of some recent private mail, which boldly claimed
> that if you implement things in that way, your language is not
> really case sensitive at all...)

So you told them about setattr() and friends, right? :)

Regards,

Martijn
-- 
History of the 20th Century: WW1, WW2, WW3?
No, WWW -- Could we be going in the right direction?