Re: [Axiom-developer] Axiom and streams

[Stephen Wilson]

C Y <address@hidden> writes:
> Not a particularly useful question, but I am curious - how serious is
> the collision conceptually between Gray Streams and Simple Streams?  Is
> one a special case of the other, for example?  (I.e. could Simple
> Streams be regarded as a special case layer on top of Gray Streams, or
> doesn't that make sense?)

I think they take different approaches, where both have advantages and
disadvantages in certain applications.  For example, I think Simple
streams have a good interface for dealing with common byte-based
operations like reading files, compressed data, encrypted bits, etc.
Gray streams take more of an abstract approach, so that a stream of
objects, say, being sent to a graphics engine for display, fits better
in the Gray streams model IMHO.

You might be able to define one in terms of the other, but then you
would not have a `true' Gray stream or Simple stream implementation,
as both define what the fundamental objects are and how they interact
with ANSI lisps stream substrate.  What would result, I think, is a
`best of both worlds' model.

Some interesting comments were made about the issues involved here:

 http://common-lisp.net/pipermail/streams-standard-discuss/2004-September.txt


Take care,
Steve