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

Re: reading NaNs

This page is part of the web mail archives of SRFI 77 from before July 7th, 2015. The new archives for SRFI 77 contain all messages, not just those from before July 7th, 2015.



Alan Watson <a.watson@xxxxxxxxxxxxxxxx> writes:

> Thomas Bushnell BSG wrote:
>> You are still pretending that "IEEE real" is a datatype instead of a
>> set of rules for computations.
>>
>>>Which of them would the syntax like "3.14" resolve to?
>> Neither, because "IEEE real" is not a datatype.
>
> When the reader reads "3.14", it must produce an object. Unless it 
> associates the string representation with that object, it must convert 
> "3.14" into an internal representation. How do you select that internal 
> representation? You will get different results if you represent it as an 
> IEEE single, IEEE double, VAX double, IBM double, some decimal flonum, 
> someextended flonum, etc.

Bingo.  

> Yes, "arithmetic" and "representation" are somewhat separate, but the 
> reader normally converts an external representation into an internal 
> representation.
>
> So, how can the reader know which representation to use?

By the user saying at load time (in one of several sensible ways,
perhaps all of which should exist) how to interpret the source file.

> (I have some answers, but here I am simply trying to clarrify the 
> question. I hope this is a help rather than a hindrance.)

It's a help indeed!