- From: Erik Wilde <dret@berkeley.edu>
- Date: Wed, 05 Aug 2015 14:40:06 -0700
- To: Data on the Web Best Practices Working Group <public-dwbp-wg@w3.org>
- CC: Annette Greiner <amgreiner@lbl.gov>
hello annette. On 2015-08-05 12:39 , Annette Greiner wrote: > You misunderstand me. When I talk about the vocabulary here, I’m not talking about the URIs themselves. I’m talking about the affordances in each response that define the relationships between the links, like the rel tags. I recognize that strict REST demands that the client discover the URIs on the fly. When I talk about hops in this context, I’m talking about the exploratory path that the client follows when following the links, not the path represented by a URI. ooops, sorry, i think your examples about having discussions about URI structure got me confused. very good then. for those affordances, if possible, it would be good to reuse existing ones and simply document their usage ("expect RFC 5005 paging links"). since the IANA registry only shows registered ones (but not proposed ones in drafts), i am maintaining a list that also includes proposed link relations: https://github.com/dret/sedola/blob/master/MD/linkrels.md if you have to mint new ones, do it, document them in the same way, and then add them to the list of link relation types used in your service. > It seems to me one would want to keep consistent URIs for product resources, since that fits with the idea of unique and stable identifiers. That seems to go against the idea that URIs should be completely changeable, but perhaps that is an exception that “goes without saying”. URIs always must be completely opaque. and imagine that you're listing products that you don't own (or not all of them). then you will have to work with whatever URI the product owner assigns, and very likely your list of product URIs will not follow one consistent pattern. the only requirement you could reasonably have is that the owner of other products would keep their URIs stable and unique, so that you can depend on them. cheers, dret. -- erik wilde | mailto:dret@berkeley.edu - tel:+1-510-2061079 | | UC Berkeley - School of Information (ISchool) | | http://dret.net/netdret http://twitter.com/dret |
Received on Wednesday, 5 August 2015 21:40:34 UTC