Hi All, We are currently revising the ALTO Protocol draft following the feedback received at IETF77. There was much helpful feedback, including a suggestion of an alternative, fully-RESTful, approach to the protocol.
One of the possible changes to make the protocol truly RESTful would be to define a single (or few) entry point for discovering resources available at the ALTO Server (i.e., services and various resources within those services); the returned document would be an index for available resources and their corresponding URLs. A previous incarnation of the protocol draft (pre-WG-document-days, see http://tools.ietf.org/html/draft-penno-alto-protocol-01#page-19) used a similar approach, but the argument was made that this made the protocol more complicated than it really needed to be. Thus, the current protocol instead follows a pattern similar to that used in many public services deployed today, where URLs are explicitly defined. There are other aspects of the protocol that might be changed to make it fully RESTful (see the thread begun by Martin Thomson: http://www.ietf.org/mail-archive/web/alto/current/msg00632.html for a more complete discussion), but the above is probably the first one which the WG should agree on. While a fully RESTful approach may provide more flexibility for the ALTO Service, but it is unclear whether this flexibility is needed in our context. In other words, we are looking for consensus and use-cases as opposed to being fully RESTful just from a compliance perspective. Thus, we are looking for input from the WG. Taking into account all previous feedback, it looks like (in our opinion) there is no consensus to change the current direction. At the current stage, we plan to update the draft to explicitly clarify the design choices that were made and avoid claiming that the service is RESTful (to avoid abusing the term). Thanks, Rich and Reinaldo _______________________________________________ alto mailing list [email protected] https://www.ietf.org/mailman/listinfo/alto
