headers
Tony Hansen | 4 Sep 00:24
Picon
Favicon

Fw: requested tweak to 2822upd grammar


I'm passing this message on to the 822 list for vetting.

Alexey Melnikov wrote:
> I realize that this might be too late now, but I've suggested this
> before and got no response. And I hope this is not controversial for
> AUTH48:
>
> Can we please redefine the msg-id in section 3.6.4:
>
> msg-id          =   [CFWS] "<" id-left "@" id-right ">" [CFWS]
>
> to be:
>
> msg-id          =   [CFWS] "<" msg-id-core ">" [CFWS]
> msg-id-core     =   id-left "@" id-right
>
> This way other documents can reference the thing inside <> more
> easily.

This change is strictly an arrangement of the ABNF and does not affect
either the intent of the document nor what's on the wire.

I've been asked to ask the list for buy in on this suggestion.

Without a definite consensus to do so, it will not be done. The RFC
Editor hasn't yet made it down to our document in their queue, so this
request is not blocking anything. Lisa has indicated a willingness to
support the change as an RFC Editor note, if there is consensus to do
so. But it *is* an extremely late request.
(Continue reading)

Permalink | Reply | 698 comments |
headers
Dave CROCKER | 4 Sep 02:48

Re: Fw: requested tweak to 2822upd grammar


Tony Hansen wrote:
>> msg-id          =   [CFWS] "<" msg-id-core ">" [CFWS]
>> msg-id-core     =   id-left "@" id-right
>>
>> This way other documents can reference the thing inside <> more
>> easily.
> 
> This change is strictly an arrangement of the ABNF and does not affect
> either the intent of the document nor what's on the wire.

Given that 2822upd has been approved and the tracker shows the document as being 
in the RFC Editor queue, I'm unclear about the impact of such a change on the 
document's processing.  Since this change does not affect semantics and has a 
minimal impact on parsing and doesn't change any terminals, and since the RFC 
Editor processing includes editing refinements of language, I would expect this 
change to have no impact on RFC Editor handling.

If these assumptions are wrong, I would prefer to skip this refinement.

However on the assumption the change can be made without having any approval or 
publication impact, I encourage the change.

Style in specifications is about human readability.  This holds true for 
meta-syntax usage too.

Over the years, I've come to appreciate the real benefit of capturing concepts 
by creating ABNF rules that might not otherwise be needed.  msg-id-core is one 
of those, as was the introduction of display-name in going from 822 to 2822.
(Continue reading)

Permalink | Reply | 1 comment |
headers
Arnt Gulbrandsen | 4 Sep 11:42
Favicon

Re: Fw: requested tweak to 2822upd grammar


Tony Hansen wrote:
> Alexey Melnikov wrote:
>> Can we please redefine the msg-id in section 3.6.4:
>>
>>  msg-id          =   [CFWS] "<" id-left "@" id-right ">" [CFWS]
>>
>>  to be:
>>
>>  msg-id          =   [CFWS] "<" msg-id-core ">" [CFWS]
>>  msg-id-core     =   id-left "@" id-right

For URL usage, what else?

> Please respond to the mailing list before Monday September 8th with a 
> +1 (for the change) or -1 (against the change).

+1. URLs matter and as Dave says, maybe there's more.

Arnt
Permalink | Reply | 0 comments |
headers
Abhijit Menon-Sen | 4 Sep 12:12
Favicon

Re: Fw: requested tweak to 2822upd grammar


At 2008-09-03 17:48:01 -0700, dhc <at> dcrocker.net wrote:
>
> However on the assumption the change can be made without having any
> approval or publication impact, I encourage the change.

Same here.

-- ams
Permalink | Reply | 0 comments |
headers
Jeff Macdonald | 4 Sep 16:14
Favicon

Re: Fw: requested tweak to 2822upd grammar


+1

On Wed, 2008-09-03 at 18:24 -0400, Tony Hansen wrote:
> I'm passing this message on to the 822 list for vetting.
> 
> Alexey Melnikov wrote:
> > I realize that this might be too late now, but I've suggested this
> > before and got no response. And I hope this is not controversial for
> > AUTH48:
> >
> > Can we please redefine the msg-id in section 3.6.4:
> >
> > msg-id          =   [CFWS] "<" id-left "@" id-right ">" [CFWS]
> >
> > to be:
> >
> > msg-id          =   [CFWS] "<" msg-id-core ">" [CFWS]
> > msg-id-core     =   id-left "@" id-right
> >
> > This way other documents can reference the thing inside <> more
> > easily.
> 
> This change is strictly an arrangement of the ABNF and does not affect
> either the intent of the document nor what's on the wire.
> 
> I've been asked to ask the list for buy in on this suggestion.
> 
> Without a definite consensus to do so, it will not be done. The RFC
> Editor hasn't yet made it down to our document in their queue, so this
(Continue reading)

Permalink | Reply | 0 comments |
headers
Frank Ellermann | 7 Sep 02:44
Picon
Picon

Re: requested tweak to 2822upd grammar


Tony Hansen wrote:

>> msg-id          =   [CFWS] "<" msg-id-core ">" [CFWS]
>> msg-id-core     =   id-left "@" id-right
[...]
> But it *is* an extremely late request.

Yes, and it could cause some confusion for two (or soon
three) NetNews RFCs in the pipeline.  I'd prefer to do
this later (= in the promotion to STD), and not as some
last minute hack with unclear consequences.

There can be only one <msg-id-core>, and it's specified
in an approved NetNews RFC at the moment.  Referenced
in an approved news URI RFC, and in the now Last Called
NetNews protocol.

IFF a future NetNews <msg-id-core> could be strictly the
same as in 2822upd I'd support this last minute hack,
but this is not the case.

BTW, the proposal to adopt the NetNews syntax wholesale
in 2822upd was proposed here, but not adopted.  At the
moment we have these relations for a msg-id-core syntax:

  (2822upd - obs) < NetNews, NetNews < (2822 - obs)
   (2822 - obs)   < 2822upd, NetNews < 3977

It would be more tricky when adding news URL relations,
(Continue reading)

Permalink | Reply | 2 comments |
headers
SM | 7 Sep 07:44

Re: Fw: requested tweak to 2822upd grammar


At 15:24 03-09-2008, Tony Hansen wrote:
>I'm passing this message on to the 822 list for vetting.
>
>Alexey Melnikov wrote:
> > I realize that this might be too late now, but I've suggested this
> > before and got no response. And I hope this is not controversial for
> > AUTH48:
> >
> > Can we please redefine the msg-id in section 3.6.4:
> >
> > msg-id          =   [CFWS] "<" id-left "@" id-right ">" [CFWS]
> >
> > to be:
> >
> > msg-id          =   [CFWS] "<" msg-id-core ">" [CFWS]
> > msg-id-core     =   id-left "@" id-right
> >
> > This way other documents can reference the thing inside <> more
> > easily.
>
>This change is strictly an arrangement of the ABNF and does not affect
>either the intent of the document nor what's on the wire.

msg-id is imported from RFC 2822 in several USEFOR and URI 
I-Ds.  Although this change does not affect the document, it might be 
viewed as controversial if it affects the other I-Ds.

Regards,
-sm
(Continue reading)

Permalink | Reply | 0 comments |
headers
Frank Ellermann | 7 Sep 15:08
Picon
Picon

Re: requested tweak to 2822upd grammar


Requested copy of an off-list mail:

> What if it were named msg-id-interior or some other
> variant instead of msg-id-core?

Well, I strongly support to have only one <msg-id-core>
concept everywhere, creating a new name for the 2822upd
variant does not exactly help with this final goal.

Are the folks wanting the 2822upd variants (sic!) sure
that they really want them ?  Or do they actually want 
some  <id-left> "@" <id-right>  excluding "obs" ?

If it is the latter, the proposed syntax doesn't exclude
"obs" and won't help them.  If it is the former they can
simply say  <id-left> "@" <id-right>  without creating
a new shorthand for this construct.

Third possibility, they actually want <msg-id-core> as
specified in RFC.ietf-usefor-usefor.  That should get
its number this year, the only MISSREF is in Last Call.

Fourth possibility, they want the future "real thing",
a <msg-id-core> limited to what both 2822upd minus obs
*and* RFC.ietf-usefor-usefor allow.

For the fourth option they should wait like everybody
else until there is an IETF consensus about the future
unification.  We are near, but not yet there.
(Continue reading)

Permalink | Reply | 0 comments |
headers
Tony Hansen | 10 Sep 01:40
Picon
Favicon

Re: Fw: requested tweak to 2822upd grammar - consensus call


Thank you everyone for posting your opinions on this matter. I'm going
to categorize them this way:

  1) fine with idea if there's no impact on rfc editor	
	Alexey Melnikov (proponent of the idea)
	Dave Crocker
	Abhijt Menon-Sen
	Lisa Dusseault

  2) fine with idea
	Arnt Gulbrandson
	Jeff Macdonald

  3) would affect other RFCs with current name
	Frank Ellermann
	SM

  4) not sure people really want this, given the way the obs stuff is
     pulled in by id-left and id-right
	Frank Ellermann

  5) fine with the idea if there're no negative impacts and consensus is
     behind it
	Pete Resnick

Given the above, I'd say that the idea has consensus behind it, but not
with the current suggested name. I'd say it's worth changing the name to
something that doesn't cause conflicts, and let people try using it. It
may be they won't be able to use it for the reasons Frank mentions, but
(Continue reading)

Permalink | Reply | 693 comments |
headers
Dave CROCKER | 10 Sep 01:51

Re: Fw: requested tweak to 2822upd grammar - consensus call


Tony Hansen wrote:
> Doing a few searches show that some names that don't cause conflicts are
> msg-id-internal, msg-id-kernel and msg-id-element. I think the choice
> should be between Pete and Lisa.

Not that you asked for additional input, but to offer a personal comment on the 
linguistic choices:  While "internal" is structurally correct, it doesn't say 
much about completeness or utility.  And "element" actually implies incompleteness.

Of the 3 choices, "kernel" nicely implies completeness and utility.

d/

--

-- 

   Dave Crocker
   Brandenburg InternetWorking
   bbiw.net
Permalink | Reply | 10 comments |

Gmane