Re: ID Tracker State Update Notice: draft-ietf-opes-ocp-core

On Thu, 2 Sep 2004, Ted Hardie wrote:

> Actually, Steve simply wanted to hold the document open for 
> discussion, since there were some things he wanted to dig through 
> further; the IESG doesn't use DEFER for that (it just continues the 
> discussion), but it was very, very early in the morning in Australia 
> when he pushed the button and he forgot.  He and I hope to circle up 
> on it next week.

That explains, and thank you for watching over the discussion.

> We did have a short discussion of the registry requirements, though, 
> and it would help IANA if we mocked up what an entry in the registry 
> would look like, so they can prepare the registry when it goes to 
> them.

I would be happy to create an example, probably using a real entry 
from OCP/HTTP profile. Where should I send it once it is discussed by 
the WG?

> Also, is there an available validator for PETDM, or is this 
> something the expert will validate?

I doubt PETDM is formal[ized] enough to automate validation. If PETDM 
happens to be useful, we probably would be able to judge whether 
further fomalization makes sense. For now, it is just a semi-formal 
mnemonic requiring human interpretation. Expert validation would be 
required.

> On the ABNF issue, if there is a different flavor of BNF which can 
> express the "n octets" rule, we can switch the reference; the IESG 
> doesn't mind folks using different flavors of notation, it just 
> needs the citations to match.  If there is no flavor of BNF that 
> matches, we can say "expressed in ABNF, except for the "n octets 
> rule"", which is a kludge, but will probably work.

I do not know of any *BNF that allows dynamic rules (other than via 
comments). Moreover, since *BNF notations vary, simply switching 
references will most likely not be possible. How about replacing the 
offending

 	data = size ":" <n>OCTET                 ; n == size

rule with ABNF compliant

 	data = size ":" *OCTET         ; exactly size octets

rule. This change can be done via Editor Note, I think. Would that 
work?

Thank you,

Alex.

> At 9:46 AM -0600 9/2/04, Alex Rousskov wrote:
>> FYI: It looks like we need to wait another IESG telechat cycle since Steve 
>> Bellovin excersized a "Defer" option to review the latest draft changes 
>> (published in May). Steve's DISCUSS is the last DISCUSS we need to clear.
>> 
>> FWIW, ID Tracker description of "Defer" says that it can only be invoked 
>> the first time the document comes up for discussion during a telechat. 
>> Since this is the second telechat for OCP Core, I am guessing that the rule 
>> applies to each revision of the document.
>> https://datatracker.ietf.org/public/pidtracker.cgi?command=view_state_desc&id=21
>> 
>> There is also a new problem recorded by Bill Fenner (without the 
>> corresponding DISCUSS): OCP Core syntax cannot be fully expressed using 
>> ABNF. ABNF cannot express an "n octets" rule where "n" is determined 
>> dynamically.
>> 
>> Since it is not a DISCUSS, I would leave the BNF "as is" because the 
>> offending rule is essentially a free-flow comment (from ABNF point of 
>> view), and we have no better way of expressing this, short of extending 
>> ABNF.
>> 
>> Alex.
>> 
>> On Thu, 2 Sep 2004, The IESG wrote:
>> 
>>> 'State Changes to IESG Evaluation - Defer from IESG Evaluation::AD 
>>> Followup by Steve Bellovin'
>>> ID Tracker URL: 
>>> https://datatracker.ietf.org/public/pidtracker.cgi?command=view_id&dTag=10380&rfc_flag=0