- From: Seaborne, Andy <andy.seaborne@hp.com>
- Date: Tue, 20 Jul 2004 17:50:00 +0100
- To: "Thompson, Bryan B." <BRYAN.B.THOMPSON@saic.com>, public-rdf-dawg@w3.org
-------- Original Message -------- > From: Thompson, Bryan B. <> > Date: 20 July 2004 17:35 > > Andy, > > Is this just the "transfer" syntax for query expressions? So, if we are > doing a bookmarkable > query, then this would be the query string? > > -bryan Yes - its the serialization of the query on the wire. To send a query from one machine to another, it has to be serialized. For a GET? based protocol, it would be (part of) the query string. With the N3QL (service-oriented view) it would be the whole thing, with the RDF Web API (model oriented), there would be a parameter with the query string. All hex encoded of course. The WG may yet decide there there are protocol parameters to be carried outside the query serialization. Andy > > -----Original Message----- > From: public-rdf-dawg-request@w3.org > [mailto:public-rdf-dawg-request@w3.org] > On Behalf Of Seaborne, Andy > Sent: Tuesday, July 20, 2004 7:00 AM > To: public-rdf-dawg@w3.org > Subject: Network Transfer Syntax > > > > An important value having a W3C recommendation in this area is to enable > interoperation of independently developed systems. For me, the fact the > name is "data access" is key - it is providing the ability of one > system to > access information on another system without needing to know the > specific > details of how the remote system is implemented. An important aspect of > this is the network transfer syntax (also known as the on-the-wire > syntax). > > The network transfer syntax (NTS) does not have to be the syntax > presented > to the application writer. This is particularly true if we have several > presentation syntaxes of a common abstract query capability (we have 3 > so > far). Each presentation syntax would be targeted at a specific > environment > or usage (they may not even be syntaxes but programmatic APIs). > > But it really does help to have one single network syntax. This > serialization of the query language should be reasonably compact, be > able to > represent the full range of query capabilities and be easy to parse to > extract the abstract query syntax. > > Local query would not need to go through the NTS, it could remain in the > application presentation syntax. > > Dan's limited complexity requirement seems most important here. While > any > presentation syntax will need to provide features that are good for its > environment (e.g. business logic formatting requirements) that does not > mean > that the NTS, the data access language, should have those capabilities > when > NTS is passing across system and administration boundaries. > > Andy
Received on Tuesday, 20 July 2004 12:52:01 UTC