Hi Toerless,

My colleague David Dong and I reviewed this, and Section 5.4 looks fine. A few 
other notes:

Section 5.2, Service Names:

We'll also need a description, assignees, and contacts for the new UDP service 
names. For brski-proxy and brski-registrar, you could tell us to replicate 
those fields from the TCP registration, but brski-registrar-rjp doesn't have a 
corresponding TCP registration to draw from.

In fact, this section also says that the "defined TXT keys column for 
brski-proxy, brski-registrar and brski-registrar-rjp for "tcp" and "udp" 
protocols are to state the following text," but it doesn't tell us to register 
a TCP service name for brski-registrar-rjp. We have registered a TCP 
"brski-reg-cmp" service name that isn't being mentioned in this section.

Finally, as David noted, service names can't be more than 15 characters long 
(RFC 6335, Section 5.1), but "brski-registrar-rjp" is 19 characters. Could 
"reg" be used in place of "registrar" instead, as in "brski-reg-cmp"?

You can see registrations that contain the string "brski" here:

https://www.iana.org/assignments/service-names-port-numbers/service-names-port-numbers.xhtml?search=brski

Minor terminology issues (no rush):

5.3: '"GeneRic Autonomic Signaling Protocol (GRASP) Parameters Registry" page' 
should be '"GeneRic Autonomic Signaling Protocol (GRASP) Parameters" registry 
group'

5.4: "registry webpage" -> "registry group"

Section references in the new registries might also be useful, but of course 
aren't required. 

Thanks again,
Amanda

On Mon Oct 20 19:24:30 2025, [email protected] wrote:
> Thanks a lot, Amanda.
> 
> I hope i did resolve all IANA issues in rev 08/09:
> 
> https://www.ietf.org/archive/id/draft-ietf-anima-brski-discovery-
> 09.html#name-iana-considerations
> 
> Would be great if you could do another round of review until IETF124,
> so
> i can then hopefully declare to have passed iANA review (and get to
> WGLC ;-)
> 
> - The RFC6690 registration procedure related text is hopefully now
> correct
>   and the draft is also an update to rfc6690 to be allowed to update
> the
>   registration procedures.
> 
> - Had a major rework of the BRSKI Discovery Parameters tables.
>   Draft is now targeting the format of one registry with multiple sub-
> registries,
>   given how i did like that representation in IANA CORE registry the
> best.
> 
> - no more multi-line registration stuff, empty cells, all tables
> should
>   still be sane if resorted by arbitrary columns.
> 
> - split into more tables so each table only has one core registration
> column.
> 
> - more precise regustration requirements for expert review for each
> tabe.
> 
> - Hopefully tables will fit into 72 characters once all drafts are
> RFC, which
>   they should be in RFC editor queue (but that's not an IANA concern).
> 
> Thanks!
>     Toerless
> 
> On Thu, Oct 16, 2025 at 10:12:25PM +0000, Amanda Baber via RT wrote:
> > Hi Toerless,
> >
> > I'm sorry, I dropped this one.
> >
> > > 1. Regarding
> > >
> > > https://www.iana.org/assignments/core-parameters/core-
> > > parameters.xhtml#rt-link-target-att-value
> > >
> > > Wrt to the first short table:
> > >
> > > Range                           Registration Procedures
> > > value starts with "core"        IETF Review
> > > all other values                Specification Required
> > >
> > > Q: This is actually not a registry table when i read the text of
> > > RFC6990, section 7.4 correctly. Right ?
> >
> > No. As far as we're concerned, this is just a list of procedures that
> > apply to different types of registration requests, presented in a
> > tabular format. You'll more often see them attached to registrations
> > of numeric values, where the "Range" label is a better fit, than
> > registrations of strings. Typically a registration procedure table
> > will look more like this:
> >
> > Range | Registration Procedures
> > 0-127 | Expert Review
> > 128-251 | First Come First Serve
> > 252-255 | Experimental Use
> >
> > > Q: if as i assume this table is not a registry (and this applies to
> > > many similar tables in other registries): Would it not be a lot
> > > cleaner
> > >  to move such a table then into the "Note" section as opposed to
> > > the place where it currently is - after the "Available Formats" ?
> > > Whereit currently sits, it looks more like part of the registry as
> > > opposed to the rules for the registry. And one may try to find (in
> > > vain)
> > > the tables entries in the CSV that follows...
> >
> > I agree. When the registration procedure is presented in a line of
> > text rather than in a table, it appears immediately beneath the
> > registry name. I just sent a note to our development team.
> >
> > > Note1: I am also asking because i was thinking i could refine the
> > > brski-discovery
> > >  ask from section 4.1.1 to extend that table to say:
> > >
> > > Range                           Registration Procedures
> > > value starts with "core"        IETF Review
> > > >> value starts with "brski"       Expert Review with Specification
> > > >> Required
> > > all other values                Specification Required
> > >
> > > But if you confirm that that table is not a registry, than such a
> > > change through brski-discovery would be an update to rfc6990 and
> > > accordingly require work with core-wg / core-parameters. So then
> > > maybe
> > > we  even discuss with them how to convert the table to a registry
> > > through a core-wg draft
> > > instead of a one-off as brski-discovery 4.1.1. now asks for.
> >
> > If this is a registry, the registry as it exists now would need to be
> > split into two (or three, pending "brski") subregistries of that
> > registry.
> >
> > A question: if you were to create a separate registry for strings
> > beginning with "brski-", while leaving this registry as it is, would
> > you then need to reserve values beginning with "brski-" in this
> > registry and state that they can't be registered here? We would still
> > need to add "value starts with 'brski*' | Reserved" to the
> > registration procedure table.
> >
> > > More inline:
> > >
> > > On Thu, Jul 24, 2025 at 08:09:02AM +0000, Amanda Baber via RT
> > > wrote:
> > > > Hi Toerless,
> > > >
> > > > Two separate issues I want to talk about.
> > > >
> > > > 1) It looks like what's being requested in Section 4.2.2 is what
> > > > we'd
> > > > call a registry with subregistries, except that the registry
> > > > would
> > > > only consist of a title ("BRSKI Variations Discovery Parameters")
> > > > and
> > > > procedures with no table, which is problematic for us. There are
> > > > a
> > > > couple of registries from ca. 2000-2005 that we've converted to
> > > > XML
> > > > in this way, but the development team that's looking at a new
> > > > presentation format is strongly discouraging this. (The beta
> > > > website
> > > > is being used to demonstrate something other than this format at
> > > > the
> > > > moment, so I can't send screenshots, but could pass on your
> > > > information to the team if you would be interested in a demo at
> > > > some
> > > > point.)
> > >
> > > Q: I hope you are referring to 4.2 and it's sub-sections and not
> > > only
> > > 4.2.2.
> >
> > Yes, it looks like I meant 4.2.
> >
> > > Q: How is section 4.2 of brski-discovery different from rfc7252 and
> > > it's
> > > resulting set of sub-registries. And that's a 2014 RFC, not 2000-
> > > 2005:
> > >
> > > https://www.iana.org/assignments/core-parameters/core-
> > > parameters.xhtml#codes
> > >
> > > The initial table in the "CoAP Codes" registry is not a registry
> > > table
> > > itself,
> > > right ? It is just shared allocation rules for the two sub-
> > > registries
> > > "CoAP Method Codes" and "CoAP Response Codes".
> >
> > So we won't create a walled-off section within a registry group,
> > within a file, unless we're creating subregistries for a parent
> > registry.
> >
> > This is a a structural issue for us, and one we have to be fairly
> > rigid about as our development team looks at registry presentation
> > design and conversion from XML to database storage.
> >
> > A table of registration procedures, for IANA, is the list of the RFC
> > 8126 or 8126-derived procedures we need to apply in order to process
> > a registration request, depending on criteria provided in a "range"
> > field or similar. https://www.iana.org/assignments/core-
> > parameters/core-parameters.xhtml#codes doesn't do this. It lists
> > ranges and associates them with a category. The subregistries for the
> > categories have the "registration procedure" field.
> >
> > An alternative presentation for the CoAP Method Codes and Response
> > Codes registries would have been to present a single Method and
> > Response Codes registry, and supply a registration procedure table at
> > the top that combined the "Method" or "Response" information in an
> > additional "Note" field.
> >
> > > Are you telling me that you would need brski-discovery section 4.2
> > > specify
> > > some non-registry rule-table so that 4.2.1, 4.2.2 and 4.2.3 could
> > > each
> > > create
> > > sub-registries in the same way as "CoAP Method Codes" and "CoAP
> > > Response Codes"
> > > do ???
> >
> > The problem is simply that we have been asked, on account of current
> > and especially future structural considerations, not to do this in a
> > registry group:
> >
> > Registry
> >
> > Section
> >
> > - Registry in section
> >
> > - Registry in section
> >
> > Whereas we can present
> >
> > Registry
> >
> > Registry
> >
> > - Subregistry
> >
> > - Subregistry
> >
> > thanks,
> > Amanda
> >
> > > > Also, a quick note: we'll also need to list the registration
> > > > procedure individually for each table/registry, no matter what
> > > > the
> > > > approach.
> > >
> > > Sure.
> > >
> > > > So a couple of suggestions:
> > > >
> > > > a) What if we were to create a new registry group at a new web
> > > > page
> > > > dedicated to this, and place individual links to all these in
> > > > https://www.iana.org/assignments/brski-parameters? I can show you
> > > > what those links look like in an existing page.
> > >
> > > Understood. I'll look into that option if the above "sub-registry"
> > > option like done for CoAP Codes falls through.
> > >
> > > > Look for "Measurement Timestamp Type" under "Registries Included
> > > > Below" at https://www.iana.org/assignments/mpls-lsp-ping-
> > > > parameters,
> > > > and you'll see a registry title and reference accompanied by a
> > > > note
> > > > that points to another location. We could do this at brski-
> > > > parameters
> > > > for each table/registry and use whatever you want for the note.
> > > > Maybe
> > > > something like "Included in the 'BRSKI Variations Discovery'
> > > > registry
> > > > group at [URL]."
> > > >
> > > > b) A little hackier, perhaps: what if, instead of creating a
> > > > section,
> > > > we incorporated the type of registry into the name? e.g., "'BRSKI
> > > > Variations Discovery: Context," "'BRSKI Variations Discovery:
> > > > Variation Type and Choices," and "BRSKI Variations Discovery:
> > > > Variations and Variation Strings"?
> > >
> > > See above.
> > >
> > > > 2) Two separate questions about table contents:
> > > >
> > > > a) I'm assuming that in Table 2, for example, "BRSKI" and "mode
> > > > vformat enroll" are meant to apply to the second and third rows
> > > > as
> > > > well as the first row. If so, would there be any issues with us
> > > > populating those empty fields? Here's an example of a registry
> > > > that
> > > > does that: https://www.iana.org/assignments/email-auth. At the
> > > > moment, I don't know whether we might need to come back later and
> > > > say
> > > > "we really need to fill in these blank spaces." I see the current
> > > > approach being more readable in the draft, so I don't think the
> > > > draft
> > > > and registry would necessarily need to match in that way.
> > > > (Another
> > > > point I just remembered here: the registries are sortable, as
> > > > you'll
> > > > see if you click on the box to the right of the column header
> > > > name in
> > > > any of those email-auth registries. This suggests to me that it
> > > > would
> > > > be better not to leave fields blank.)
> > >
> > > Right. I will add note to next-rev that the registry can/will have
> > > the
> > > fields filled and that not-filling them is just for nicer editorial
> > > visibility.
> > >
> > > > b) The above would make more sense for the CSV file associated
> > > > with
> > > > each registry and the JSON file we're going to associate (this is
> > > > what you can see in registries linked from
> > > > https://beta.iana.org/protocols right now). But for that reason,
> > > > separating "mode," "vformat," and "enroll" with commas would also
> > > > likely make more sense than line breaks.
> > >
> > > Right. Next rev.
> > >
> > > Thanks a lot!
> > >     Toerless
> >
> > Many apologies,
> > Amanda

_______________________________________________
Anima mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to