Re: conneg considered harmful
"Eric J. Bowman" <[email protected]>
| Newsgroups | gmane.comp.web.services.rest |
|---|---|
| Organization | Bison Systems Corporation |
| Message-ID | <[email protected]> |
Mark Baker wrote: > > There are multiple forms of conneg and I think Roy has been pretty > clear about his disdain for server driven. Larry Masinter even > apologized in recent years for pushing for it back in the day. > Would be nice to have some links, such discussions have eluded me. My apologies for overreacting, I took Roy's meaning as an argument against the late binding of representation to resource altogether, as I lacked such context. > > Besides, client driven uses hypermedia so is way more RESTful. > Adding a round-trip, unless there's some way to make user preferences "sticky" for subsequent requests (like negotiation headers). Leading to sessions, at least as far as HTTP 1.1's concerned, so "more RESTful" is debatable as to real-world implementation. But, I see where other protocols could do things differently, avoiding such results. -Eric