[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

syntax coloring & type inference [Re: Common Lisp "let" / Re: Curl, ...]

It struck me that having the
different pairs of parens is
a sort of poor man's syntax
coloring (especially in the
case of setting off comma-ed
forms in backquoted lists).

The use of type inference to
quickly check if "what I said" 
was clearly not "what I meant"
has a similar flavor.

(also like the use of a cross
listing file, to find out if:

	DO 10 I = 1.25

codes a loop or an assignment)

I guess this is a matter of
tools rather than language,
but it seems that feedback
resolving ambiguity may be
as important as reducing it
in the first place.



>                  If your code didn't support the spec -- maybe your
> push/pop/top implementation satisfies the *type*, but *implements* a queue,
> not a stack -- then it was rejected by the compiler.

This sounds like the idea that,
after being changed, a module
must pass the test suite that
the previous version did.  (if
it is a bug fix, it should pass
the old suite, plus additional
tests that tickle the bug)

This is common practice; it's
just not done by the compiler.