Thomas Bushnell BSG <xxxxxx@becket.net> writes:

>>> It is an explanation of why your appeal to tradition is ill-founded;
>>> the tradition actually points the opposite way.
>>
>> There is no tradition in not providing a read syntax for "error
>> objects".
>
> Really?  Can you give me some pointers to the Lisp systems which do
> provide read syntax for all the error objects they have?

Could you show an object which does not have read syntax in Lisp
*because* it's an error object?

Common Lisp condition objects, if you mean them. are not comparable to
NaN. A condition is caught out of band; a NaN is returned instead of
a more exact result and gets printed along with the normal way of
printing results.

When one part of the program prints numeric results to a file, and
another part later reads them, what is the point in breaking this
communication channel for NaNs?

Sigh. With this attitude I have no hope in getting a decent portable
numeric system from Scheme. Implementations are on their own. Time to
move to more sane languages.

--
   __("<         Marcin Kowalczyk
   \__/       xxxxxx@knm.org.pl
    ^^     http://qrnik.knm.org.pl/~qrczak/