Re: Translating from a VERY "loose" language

[Rici Lake <lua@...>]

On 11-Jun-07, at 4:19 PM, Steve Heller wrote:

> I suppose so. But then perhaps there should be
> different operators that behave like "<=", etc. with
> respect to types, but for (in-)equality of values.
> Maybe === and ~==? That would simplify problems like
> the one I have.
>
> Or am I (almost) the only one with this problem? If
> so, then of course it doesn't make sense to change the
> language to meet my peculiar needs.

Certainly there is an argument for the <> operator,
meaning "orders less or orders greater", which would
differ from ~= in that it would throw an error if
the arguments have no ordering relationship. (Consequently,
it would throw an error if you tried 2.0 <> NaN).

I doubt whether there would be enough use of it to
justify the (small) implementation cost, or the
(rather larger) documentation cost.

IEEE-745 contemplates a rather large collection of
ordering operations; few languages actually implement
them, though. As another datapoint, it is not uncommon
for languages to implement two or more "equality"
operators -- and the difference between these operators
often figures prominently in FAQ lists :)