Re: [apps-discuss] draft-ietf-weirds-bootstrap-00 and our lawn -- feedback?

On 12 Feb 2014, at 3:59 am, Andrew Newton <andy@hxr.us> wrote:

> Mark,
> 
> Thank you for looking into this and for being proactive.
> 
> With regard to your idea to use templates in the registry, it seems to
> me the current specification is already doing an implicit template of
> the sort http://example.com/blah/{rest-of-weirds-uri}. As one of the
> participants of the WEIRDS effort, I think your idea is worth
> consideration. While the group has looked at templates but not found
> favor with them, I believe what you are suggesting does not have the
> issue to which the group objected to previously.

OK. To be clear, it was just a suggestion -- there are a lot of ways to do this, but templates seemed like a way to get there without inventing too much yourself (which seems to be the point, since you're already reusing HTTP :).

> As for baked URIs, there was a suggestion to use .well-known with
> which the WEIRDS group is struggling (very interrelated to
> bootstrapping). Would you still suggest the use of .well-known with a
> registry? Is the use of .well-known not just to avoid collisions but
> to also allow applications to identify URIs by application type
> without the need for creating a new URI scheme?

It really depends on the use case, which I don't feel like I yet have enough grasp of. I don't want to position .well-known as a panacea, because it's not -- it's for very specific circumstances (basically, when you need to bootstrap from a hostname, not a URI).

Cheers,

> 
> -andy
> 
> On Tue, Feb 11, 2014 at 2:03 AM, Mark Nottingham <mnot@mnot.net> wrote:
>> I've been in an on-again, off-again discussion with some of the WEIRDS folks about their work and its relationship to <http://tools.ietf.org/html/draft-ietf-appsawg-uri-get-off-my-lawn>.
>> 
>> That seemed to precipitate <http://tools.ietf.org/html/draft-ietf-weirds-bootstrap-00>, which describes a way to "bootstrap" a WEIRDS interaction, and I've been asked for feedback.
>> 
>> Reading through it, I have some issues, but I since uri-get-off-my-lawn is a product of this WG, not just my draft, I'd rather get a sense of what the WG thinks overall, rather than just giving my own feedback.
>> 
>> AIUI, the bootstrap draft uses a number of IANA registries to contain a mapping of (domain names, IPv4 networks, IPv6 networks, AS numbers) to URLs for their RDAP services.
>> 
>> Those URLs are then used as base URIs for further interactions; for example, if example.com had a registry value of "http://example.com/lookup", you'd look up "foo.example.com" as "http://example.com/lookup/domain/foo.example.com".
>> 
>> To me, this seems better than the previous solution (where you assumed that example.com had something available at a certain path), but it still "bakes" URLs into the spec, relative to that base URI. I.e., the "/domain/whatever" bit above is locked into the spec and unchangeable, AIUI.
>> 
>> So, while they avoid collisions (probably), they still risk the other problems that the "get off my lawn" draft cautions against, AFAICT.
>> 
>> If I were doing this protocol and I still wanted to use a registry (questionable IMHO), I'd allow each entry to contain a set of URL templates, identified by link relations, that allows a one-step lookup without baking in any URLs.
>> 
>> E.g.,
>> 
>> domain: example.com
>>    rel: domainlookup    href-template: http://example.com/lookup/{domain}
>> 
>> I'm very curious to hear what other APPS folks think about this -- especially those of a Web bent. We're trying to line up some conversations about this in London, and I'd like to inform them with the WG's perspective, rather than just my own.
>> 
>> Cheers,
>> 
>> --
>> Mark Nottingham   http://www.mnot.net/
>> 
>> 
>> 
>> _______________________________________________
>> apps-discuss mailing list
>> apps-discuss@ietf.org
>> https://www.ietf.org/mailman/listinfo/apps-discuss

--
Mark Nottingham   http://www.mnot.net/