[newtrk] SRD - normative or not?

[John Leslie <[e-mail]>]

John C Klensin <[e-mail]> wrote:
> 
> I have to agree with Larry.  There is nothing inherent in the
> ISD proposal that requires that it contain either more or less
> information, or a different granularity, etc., from SRDs.

   I hesitate to comment on ISDs -- I'm not sure we're all talking
about the same thing.

   However, I'm happy to talk about SRDs.

> As I understand things, there were, when you proposed SRDs, only
> two key differences between the ISD concept and the SRD one:
> 
> (i) As mentioned above, ISDs are explicitly normative,
>     to the point that they become the formal definition of
>     the standard, are the real subject of IETF Last Calls
>     with RFCs being included in the last call as normatively
>     cited supporting documents.  SRDs, as I understand them,
>     do not have that property.

   SRDs are not normative, IMHO. They describe sets of RFCs, of
which each RFC may or may not be normative. If this is not clear
in the draft

http://www.ietf.org/internet-drafts/draft-otis-newtrk-rfc-set-02.txt

feel free to propose text to make it clearer.

> (ii) Largely because they have that normative property,
>      ISDs must contain at least some text, and normative text
>      at that, explaining what they are about.  They cannot be
>      just list of things that go together, providing
>      informative cross-references among those things.  When
>      the SRD concept was first floated, I understood SRDs as
>      potentially being links-only, with no local substantive
>      text.  As of the time the WG meeting in Paris ended, I
>      understood that SRDs had evolved to potentially contain
>      some local text and possibly even to require some. 

   Admittedly, I was not _in_ Paris, but I was following the audio.
I do not recall anything about requiring text, least of all any
normative text, in SRDs.

   The current I-D makes it clear that <srdref> is an empty element,
which _cannot_ contain text. I suppose it could be argued that the
<core>, <extensions>, <guidance>, <replaces>, <experimental>, and 
<companion> sections could contain text, although that is not the
intent.

   The only free-format sections are intended to be <title> and 
<description>. <title> is intended for searching, and <description>
is intended merely to describe the set, not its members.

> So there may be little substantive difference between the two
> proposals along that text versus no-text dimension any more.

   IMHO, SRDs do not contain any normative text. I accept John 
Klensin's opinion that ISDs _must_ contain normative text. I think
this difference remains unchanged.

--
John Leslie <[e-mail]>
.
newtrk resources:_____________________________________________________
web user interface: http://darkwing.uoregon.edu/~llynch/newtrk.html
mhonarc archive: http://darkwing.uoregon.edu/~llynch/newtrk/index.html