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

SRFI 21 must be a pure extension of SRFI 18

For the record I want to repeat here what I said on the SRFI 18
discussion mailing list.


> Date: Thu, 11 May 2000 22:06:35 -0400
> X-Authentication-Warning: trex.IRO.UMontreal.CA: feeley set sender to feeley@xxxxxxxxxxxxxxxxxxxxx using -f
> From: Marc Feeley <feeley@xxxxxxxxxxxxxxxx>
> To: srfi-18@xxxxxxxxxxxxxxxxx
> Subject: real-time multithreading SRFI
> Resent-From: srfi-18@xxxxxxxxxxxxxxxxx
> X-Mailing-List: <srfi-18@xxxxxxxxxxxxxxxxx> archive/latest/70
> X-Loop: srfi-18@xxxxxxxxxxxxxxxxx
> Precedence: list
> Resent-Sender: srfi-18-request@xxxxxxxxxxxxxxxxx
> I have just asked Michael Sperber to publish my real-time extension to
> SRFI 18 (it will probably be SRFI 21, available tomorrow).
> Let me explain why I'm doing this.  I would like SRFI 18 and SRFI 21
> to be entirely compatible (i.e. an implementation that supports SRFI
> 21 can also claim to support SRFI 18).  The problem is that over the
> discussion period for SRFI 21 there might be some changes to the
> specification of SRFI 21 that require similar changes to SRFI 18, and
> that would not be possible if SRFI 18 is finalized before SRFI 21.
> So I ask everyone interested in SRFI 18 to temporarily move the
> discussion to SRFI 21, and once it is finalized (or very close to it),
> we can come back to SRFI 18.  I think both SRFIs have a place, because
> not all Scheme implementors will be willing to accept the added
> complexity and run time overhead of the real-time thread system.
> I will have more to say about SRFI 21 when it is put on the SRFI
> web page.
> Marc