Re: [apps-discuss] Requirement for "obsoletes" in Abstracts

[Ned Freed <[e-mail]>]

> > I just posted a note to the RFC Editor suggesting that all this points to
> > some benefit in reviewing and revising the RFC Style guide.

> Agreed; and then we need to reflect those changes into the I-D checklist.

> > From that perspective, what it obsoletes isn't useful in the
> > Abstract, especially since Abstracts are supposed to be information-dense.

> It's particularly not useful because the abstract is, at least with
> the current template, on the same page as the header information,
> which already contains the obsoletes/updates information.

Well, the theory, as I understand it, is that the Abstract is actually a
separable entity designed to be extracted and use elsewhered. Therefore it
needs to be self-contained, to the point where it duplicates stuff that's on
the same page.

There are two problems with this theory, however. The first is that separable
Abstracts make all sorts of sense when the full document is not readily
accessible (only in print journal, behind a paywall, blah blah blah.). But
that's not the case for RFCs. It might have made sense even for RFCs when data
transfer was expensive, but these days pulling out the Abstract is more trouble
than it's worth - why not just send the whole thing? The only remaining
use-case I can think would be some sort of index, and any such index done
competently will include much more extensive forward and backwards pointers of
its own. (And it is axiomatic that the far more important forward pointers, 
that is, "this document is obsoleted/updated by RFC N", cannot possibly appear
in the Abstract.) 

The second problem is really an extension of my point about forward pointers.
Simple backwards pointers just don't give enough context to justify cluttering
what should be a straightforward statement of what *this* document does..

> To my mind,
> that makes it 100% useless to include that in the abstract.  I would
> support a recommendation that specs SHOULD explain in the introduction
> *why* this document obsoletes another if the reason is not blazingly
> obvious (obvious as when, for example, RFC 5321 obsoleted RFC 2821).

Agreed. And in the process it adds the context these simple statements in the
Abstract lack.

				Ned