[r6rs-discuss] [Formal] Bytes<->vector procedures

From: AndrevanTonder <andre>
Date: Tue Nov 14 19:45:39 2006

On Tue, 14 Nov 2006, Abdulaziz Ghuloum wrote:

> Sorry, I don't get it. What does efficiency of implementation have anything
> to do with inclusion in R6RS?

To answer your question, I am not sure what efficiency has to do with inclusion.
I did check that "efficient" and its derivatives occur in 15 places in the
document.

I do think, though, that usefulness should play a role. How often does one
need to operate on lists of numbers, as opposed to vectors of numbers? In my
experience, not very often. For this reason, I believe that bytes<->vector
operations would be more useful than the current bytes<->list operations, and I
am surprised that the latter was chosen over the former.

> I can implement these in many implementations
> efficiently already.

The bytes->vector operations are essentially fancy memory copy operations,
which can be microcoded if they are primitive.

Andre
Received on Tue Nov 14 2006 - 19:41:38 UTC