Hi all,
Il 22/02/2019 14:41, Gavin Brown ha scritto:
Thanks Andy, Mario and Bernhard for the useful feedback.
It seems to me that there is a lot of appetite for a way to replace
jCard: in addition to F2F conversations and discussions in the ICANN
world, I've also had private feedback on this draft which indicates a
lot of frustration with jCard from implementers.
If anyone implementing an RDAP client or server has some experiences
with jCard that they can share that would be very helpful in
understanding the demand for an alternative.
jCard was added in draft-ietf-weirds-json-response-03, so the last
version to have "native" contacts was -02.
In it, entities in DNRs had a different syntax to entities in RIRs,
however, there was a lot of similarity. Adding jCard helped in that the
contact information got moved into a subordinate object property rather
than being mixed in with the metadata (handle, roles, links, etc).
You can use the syntax for the contact info in the -02 draft and replace
the "eppContactInfo" object in my draft with a generic "contactInfo"
object such as the following:
{
"objectClassName": "entity",
"handle": "XXXX",
"contactInfo": {
"entityNames": [
"Joe Bob, Inc.",
"Bobby Joe Shopping"
],
"postalAddress": [
"123 Maple Ave",
"Suite 90001",
"Vancouver",
"BC",
"12393"
],
"emails": [
"[email protected]",
"[email protected]"
],
"phones": {
"office": [
"1-958-555-4321",
"1-958-555-4322"
],
"fax": [
"1-958-555-4323"
],
"mobile": [
"1-958-555-4324"
]
}
},
// rest of metadata
}
G.
I would like to give my contribution to the discussion. In my opinion,
the real mess of jCard is the massive use of jagged arrays which makes
deserialization harder than having objects with fixed members.
On the other side, I think jCard provides a very flexible representation
of entity information. Basically, it is a list of 4-tuples where the
fourth item is either simple or structured value and the addition of a
new property results in defining a new value for the first item.
That being said, the possible representation for an RDAP entity coming
in my mind are the following:
1) A structure including a set of fixed members each having a specific
meaning (like the one proposed by Gavin).
*Pros*: Easy deserialization. Conceptually elegant. *Cons*: Not very
flexible. Any additional information provided by a server will not be
recognized in deserialization. If we consider the 80-20 ratio, 80% of
most common information will be recognized, 20% of uncommon information
will not.
2) A JSON transformation of vCard different than jCard. According to
that, the contact information could be represented as an array of
objects including a set of fixed members having a generic meaning like
the following:
{ "key" : "string",
"simple" : "string",
"structured": [
{"key": "string",
"simple": "string"}
]
}
In brief, something more generic and, therefore, more flexible but, at
the same time, simple to be deserialized. Obviously, "simple" and
"structured" are alternatives.
*Pros*: Very flexible***Cons*: Not very conceptually elegant. The
meaning of each contact property is specified by a string value (the
same as it happens in jCard)
3) A hybrid solution. Most common properties are modelled by objects
whose structure is fixed (e.g. "postalAddress" could be on object
containing "street", "city", "cc" and so on) like those proposed in the
first solution, while the possible remaining properties could be
represented as an array of the same item as described by the second
solution. No information would be lost in deserialization.
Hope this could be useful.
Regards,
mario
On 19/02/2019 11:23, Andy Newton wrote:
On Tue, Feb 19, 2019 at 08:05:50AM +0100, Mario Loffredo wrote:
Hi Gavin,
if I understand correctly, this extension involves only those RDAP entities
in common with EPP (i.e registrant, admin, tech, billing), doesn't it?
If so, what about the other entities (e.g. registrar, reseller) ? Should
they be represented by jCard ?
I have the same question and concern. While I support moving away
from jCard, we should not be focusing only on EPP especially since RDAP was
first widely adopted and is used today by non-EPP communities. RDAP itself is
not a product of the EPP community but was an outgrowth of experiments
conducted by the RIRs.
That said, being compatible with EPP is certainly a requirement in my opinion.
Given this is a rather substantial change, we should also be thorough in our
approach:
1. We should dig up the pre-jCard RDAP drafts and see if there is good stuff
there.
2. We should either consult or repeat the work of CN-NIC during the WEIRDS
working group where they study the Whois output of all available registries and
found what was needed.
3. We should also pay attention to the discussions around contacts in JMAP
now going on in the IETF.
Overall, what is specified here looks good to me. Here are my comments for
improvements:
1. The country and region codes should be tied to ISO-3166 or a superset.
2. There should be a place to spell out both region and country. Some
registries do not collect 3166 codes.
3. Phone, fax, email should be arrays because some registries collect
multiples of these.
4. There should be an indicator noting that the contact information is for an
individual.
-andy
_______________________________________________
regext mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/regext
--
Dr. Mario Loffredo
Servizi Internet e Sviluppo Tecnologico
CNR - Istituto di Informatica e Telematica
via G. Moruzzi 1, I-56124 PISA, Italy
E-Mail: [email protected]
Phone: +39.0503153497
Web: http://www.iit.cnr.it/mario.loffredo
_______________________________________________
regext mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/regext
