From: Henry Spencer (henry@spsystems.net)
Date: Thu Jun 06 2002 - 14:20:39 CDT
On Thu, 6 Jun 2002, Forrest J. Cavalier III wrote:
> How is this draft going to be relevant? Unless implementors
> are of the perception that life will be better after
> implementation, then they aren't going along with it.
This is true of any standard. Compliance is likely to be piecemeal at
first, and always incomplete because some people just don't get it.
> It surely would be easier for implmenters if the present
> draft were split:
> - USEFOR Non-English character sets.
> - USEFOR Header extension methods
> - USEFOR Interoperation with mail and archives
> - USEFOR MIME Encapsulation
> - USEFOR Verification, Security, and Privacy
> - USEFOR Server control messages
First, a point of curiosity: where, in that, is the basic article format?
You know, what this whole thing is supposed to be about?
More substantively, we've discussed this before, and the conclusion has
always been that splitting the draft up is a lot of extra work with very
little payoff. It makes the overall structure harder to grasp -- that way
there is no central table of contents, and the individual documents may
get updated independently, so you have to determine the latest version of
each -- and an implementor still has to read all of it, if only to make
sure that item X doesn't affect him.
> We have a document that fits together, so splitting it
> all apart would not generate inconsistencies and conflicts.
Not at first.
> Then implementors could say that they are compliant with
> individual RFCs...
And someone who just wants to know whether his news system is up to spec
has to compare against the list of RFC numbers. And for N RFCs, there are
2^N-1 different "compliant" possibilities. This is *not* an advantage.
Indeed, it is a defect. Bad idea.
> ...instead of saying "Bah! who needs all
> that stuff?! I'm not aiming for compliance."
People who would say that... will say it either way.
Henry Spencer
henry@spsystems.net