[r6rs-discuss] [Formal] bytevector aliasing severely impedes optimizations

From: Ludovic Courtès <ludovic.courtes>
Date: Mon Mar 5 08:17:46 2007

Hi,

Bradley Lucier <lucier_at_math.purdue.edu> writes:

> It
> could be said by analogy that the proposed R6RS libraries offer
> *only* a (char*) (more tamed than in C), which solves one small
> class of problems (how to allow semi-portable, low-level translation
> between data types that can be considered sequences of bytes; how to
> write I/O device drivers; ...) while completely ignoring a much
> larger and more important class of problems (allowing fast and memory-
> efficient access to large arrays of homogeneous numerical data).

Personally, I don't think of bytevectors as an SRFI-4-like type of
homogeneous vectors, in which case the concerns you raise would be
valid. Instead, I really consider them as a "portable `char *'",
designed to facilitate binary I/O---although I do not know for sure
whether this was the authors' intent.

If this is the case, the reasons why non-aliasing is not required for
`char *' in C99 (as you said) would probably also apply to bytevectors.

Thanks,
Ludovic.
Received on Mon Mar 05 2007 - 08:17:37 UTC