ietf-nntp draft-ietf-nntpext-streaming-00

[Jeffrey M. Vinocur]

On Wed, 4 Jun 2003, Ken Murchison wrote:

> Jeffrey M. Vinocur wrote:
>
> > In particular, I was copying the TAKETHIS examples from the IHAVE ones in
> > the main draft, and I had to collapse three of the IHAVE examples (435,
> > 436, 437) into one 439 example.
> 
> Well, not really.  If you think about it, CHECK/TAKETHIS is just a split 
> version of IHAVE.  The functionality and response codes of 
> CHECK+TAKETHIS add up to IHAVE.  This is exactly how I have implemented 
> these in Cyrus.

Indeed, indeed.  I'm thinking about the optimization where TAKETHIS may be
used without an initial CHECK.  In that case, knowing the difference
between refusal and rejection is definitely useful.  (If a nontrivial
fraction of articles are being refused, then you want to switch back to
using CHECK.  But if very few articles are being refused, then regardless
of how many are being rejected, using TAKETHIS alone is optimal.)

I mean, sure, what we have now is clearly fine, and I'm not arguing to 
redo the whole thing to address this.  But if we *were* starting from 
scratch, I'd want to make these changes.

-- 
Jeffrey M. Vinocur
jeff at litech.org