[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.



Marcin 'Qrczak' Kowalczyk <qrczak@xxxxxxxxxx> writes:

> Thomas Bushnell BSG <tb@xxxxxxxxxx> writes:
>
>>> The programmer is a user of IEEE exposed by Scheme, and expects
>>> that the Scheme implementation will provide an IEEE-conformant
>>> environment. If it doesn't, it's less useful.
>>
>> The user expects that?  Really?  All of them?
>
> No, only those who care about the precise semantics of inexact reals.

Wrong.  There may actually be users who care about this which would
like something other than IEEE.  It does happen, you know.

>> How about the user expects that, if he has requested an
>> IEEE-conformant environment, he will get it?
>
> Managing multiple sets of meanings of the same constructs is
> problematic conceptually and makes cooperation between libraries
> harder. Would it apply lexically or dynamically? Could I have multiple
> interpretations in the same program? If yes, how would passing values
> between the two interpretations work, i.e. are they constrained to the
> same representation, or must data be somehow converted between them?
> If not, how can I combine two libraries which assume different
> interpretations?

Yes, because there are questions to be answered, we should just make
an obviously wrong answer do for everything!