PEP 3107 and stronger typing (note: probably a newbie question)

Bruno Desthuilliers bdesth.quelquechose at free.quelquepart.fr
Sun Jul 1 02:45:25 CEST 2007


Alex Martelli a écrit :
> Bruno Desthuilliers <bdesth.quelquechose at free.quelquepart.fr> wrote:
>    ...
> 
>>I still maintain that the primary *practical* reason behind static 
>>typing is to provide optimization clues for the compiler. You can (or at
> 
> 
> It's definitely a helpful aspect, yes -- given that compilers able to
> infer types are still not very common for widely used languages;

or given that languages using a type-inferenced based system are not 
very widely used...

(snip)

>>least could) have declarative static typing with very few type 
>>*checking* - I may be wrong here but I think one could even write a C
>>compiler without *any* type checking. Heck, the programmer said it's a
>>char*, so it must be one, right ?-)
> 
> 
> That compiler, I believe, would violate the ISO standard for C (so
> calling it a "C compiler" would be about as correct as calling it a
> banana, in my view:-).

Lol. That being said, there were bananas - oops, I meant C compilers - 
before the standard existed !-)

> 
>>wrt/ proofs of correctness, I'll just point to the spectacular failure
>>of Ariane, which was caused by a *runtime* type error in a system 
> 
> 
> I like a quote by Knuth -- "beware this program may have bugs as I have
> only proven it and not tested it":-)

Yeps.

> 
>>Hmmm... For a dinausor, C seems well alive. Can you remind me which 
>  
> So do chickens.
> 
I'm afraid I didn't get the joke... Are you saying that C is a rather, 
well, primitive language ?



More information about the Python-list mailing list