- From: Tyson Chihaya <Tyson@DataChannel.com>
- Date: Mon, 15 Nov 1999 16:46:02 -0800
- To: "'Sergey Melnik'" <melnik@DB.Stanford.EDU>, caro@Adobe.COM
- Cc: RDF Interest Group <www-rdf-interest@w3.org>
To take the opposite side for a moment, HTTP IS meant to be an extensible protocol. I generally don't advocate overloading the semantics of PUT, GET, POST, whatever, but I do think that introducing new verbs for doing the sorts of things that you mentioned is a viable solution, if the protocol you have in mind is similar to the HTTP model (ie: stateless and connectionless). Just as WebDAV has extended the HTTP/1.1 protocol by providing additional request methods (PROPPATCH, PROPFIND, LOCK, etc), we can likewise extend the HTTP method set for our own specific use, and embed XML, RDF or whatever in the request body (and even specify a particular schema for the XML/RDF). What I think you're also saying is that you would like to see an extensible, self-describing generic protocol that only uses serialized RDF models for describing the request (and no request URI/header type syntax to wrap it). I think that this would be nice (and a definite direction for the future), but it's also nice to be able to extend a model that is already in place (ie: HTTP). Another nice thing about HTTP is that headers can be quickly and easily parsed by proxies and firewalls. --tyson -----Original Message----- From: Sergey Melnik [mailto:melnik@DB.Stanford.EDU] Sent: Monday, November 15, 1999 1:14 PM To: caro@Adobe.COM Cc: RDF Interest Group Subject: Re: RDF manipulation protocol? "Perry A. Caro" wrote: > > Will Sargent wrote: > > I've been working on integrating bookmarks in Mozilla with a remote > > database, so that users can have collaborative bookmarking. > ... > >I've gone with the second approach for now, but it strikes > > me that the last thing the world needs is another hack protocol, and I'd be > > interested in a better solution. > > I would suggest using WebDAV (though one might argue that it too is another > hack protocol :-), which provides you with protocol for PUT, GET, DELETE, > etc. of URLs, with locking. You'll have to figure out how to map your > bookmark datastructure into some hybrid of a URL tree and RDF data in WebDAV > properties, but it seems feasible, and since the code is free and does 80% > of what you want, its worth considering. In my view, WebDAV-like approaches do not provide a future-proof solution. I advocate for using a generic low-level transport mechanism that allows exchanging RDF models as messages carrying well-defined semantics. One advantage of such solution is extensibility. Furthermore, if you need additional operations like "notify", "query", whatever, you are not forced to fit them into PUT, GET etc. For more reasons, look at GINF. Sergey
Received on Monday, 15 November 1999 19:55:26 UTC