Re: LP naming convention
Patrick Mevzek <[email protected]>
| Newsgroups | gmane.ietf.provreg |
|---|---|
| Organization | Dot And Co |
| Message-ID | <[email protected]> |
Wil Tan <[email protected]> 2011-04-15 14:12 > James Gould had voiced out a valid point, i.e. the elements are named > in a lowercase-with-underscores rather than camelCase. The latter is > more consistent with the rest of EPP, so I propose updating it to the > following on the next iteration of the draft: > > application_id => applicationID > trademark_name => trademarkName > trademark_number => trademarkNumber > trademark_locality => trademarkLocality > trademark_entitlement => trademarkEntitlement > pvrc => PVRC (or stay all lowercase, I'm ambivalent) > > Any comments or objections? I agree very much with this change, even if it does not change anything technically I think it is better to stay in line with the other EPP drafts, and I wanted to suggest this but feared to be too much a nitpicker, so thanks James to have raised the point :-) On a related note, for me the 4 trademark nodes are atomic, as reprensenting a single "entity". So for me it would make even more sens to have a <trademark> node with the name, number, locality and entitlement being children of it. I think it makes things clearer (as to see what is related to the trademark per se, and what is related to the trademark use in a given registry startup phase), and easier to expand if needed later on. I have no direct experiences with trademark claring houses, but I agree their input here would be very much valuable. Also seeing what was done for ENUM as Bernie suggested could be useful. As for the the draft, some quick comments: - §2 : Maybe add a not that an applicationID is like a domain ROID, the registry should make sure it is a unique number, which is needed later on (see §3.2 comment below) - §2.1 : I think the element is always useful, and it should not be derived from the time or the applicationID encoding. It can of course not be a MUST, but I would think having its presence as a SHOULD would be best. - §2.2 : for <lp:status> I would advise using a scheme similar to <domain:status> such as having the string as an attribute so that the node data is optional but could be a text in human form with more info. In order not to repeat problems with have with EPP, the draft should clearly specify if and how registry can use other statuses (than those defined in the draft) if they need. The schema should be in such a way that adding new statuses is easy, and can be done without breaking everything. Also, yes, I think we should allow multiple statuses at a time. - §2.3 I think a date is missing, either (trademark) application date or granted date, or both. - §3.2 when the registry does a lp:info, the applicationID should be enough, as soon as the registry makes it unique (same trademark used in 2 phases would have 2 applicationIDs) The lp:phase should of course be in the registry reply. In the registry reply I believe it to be strange that all nodes are optional, except the applicationID. If an applicationID exists it means something has been submitted so not all elements later can be optional, at least one or some must be defined. - §3.2.1 I fail to parse the first sentence : « The client MUST ensure that any successful <info> command results in a response that an <lp:infData> element is returned in the response. » Maybe it is completely clear for English speaker, in which case, can you make it just a little simpler to non native English speaker ? - §3.3 we take the case that 1 applicationID = 1 domain name Are there cases where we should or can consider 1 applicationID = more than one domain names ? (either because multiple domain names can be derived from the same trademark depending for example on canonicalisation rules, or because the trademark owner gives a list of domain names ordered from most prefered to least prefered, in case its first choice is alreay taken) Same remarks as before on all elements being optional: I think at least one or some may be needed. I'm uneasy with the registry reply, specifically the domain:creDate and domain:exDate and also the 1000 result code. Since there is further processing, I think we should have a 1001, not a 1000. Also, since the domain name is not really registered, the dates will be wrong. I know that at least one of them is mandatory, but I think in this way this create a problem. Registrars will need to adapt their system. There should be also some text about the domain:period asked by the registrar: in some launch phases, the registry may mandate the initial registration period, so some text should make clear what happens when the registrar asks for another period than the one allowed by the registry (is the operation accepted or refused ?) - §3.3.2 : same problem for first sentence as §3.2.1 above - §3.4 : same comment as above, the applicationID is enough in my view I also think that the update section should more concentrate on updating the info related to the application instead of the domain name. During the launch phases typically domain names are not published, so there is no sens changing their nameservers. Also for the contacts/registrants there may be registry policies forbidding change because, for example, the registrant may be the trademark owner or something like that. On the contrary one may need to change data about the application, such as adding the application granted date if it happens after the initial submission, or add a new preverification code. - §3.5 : as above, isn't the applicationID enough ? Hope that helps, -- Patrick Mevzek _______________________________________________ provreg mailing list [email protected] https://www.ietf.org/mailman/listinfo/provreg