DanCon: oh phtpht. I thought it was an XML schema for their DAML/RDF stuff. But they've gone and done a separate XML syntax.
jonb-gk: extends the RELAXNG schema for RELAXNG to define an RDF AS language
jonb-gk: possibly useful to define RDF extension languages in
jonb-gk: for example some noodling on http://www.openhealth.org/RDF/RDFSKIF.html a KIF like language built on the RDF abstract syntax
jonb-gk: err that would be ...
jonb-gk: some noodling on KIFish a KIF like RDF extension language
jonb-gk: possibly useful to define RDF extension languages in
jonb-gk: for example some noodling on http://www.openhealth.org/RDF/RDFSKIF.html a KIF like language built on the RDF abstract syntax
jonb-gk: err that would be ...
jonb-gk: some noodling on KIFish a KIF like RDF extension language
DanCon: nifty inference rule markup in the XML version and stylesheet
DanCon: cf spec-prod thread on inference rule markup
DanCon: cf spec-prod thread on inference rule markup
DanCon: by Leigh Dodds June 06, 2001
DanCon: Jelliffe, Ogbuji, et. al. on modeling, schemas, simplicity, etc.
DanCon: tangentially related: I noticed the XML Schema REC has a nice UML-ish diagram... I'd like to reproduce that diagram using the schema-for-schemas, via XSLT, to RDF, then thru my rdf->dot->SVG tools.
DanCon: found this article from the relax-ng TC home
DanCon: Jelliffe, Ogbuji, et. al. on modeling, schemas, simplicity, etc.
DanCon: tangentially related: I noticed the XML Schema REC has a nice UML-ish diagram... I'd like to reproduce that diagram using the schema-for-schemas, via XSLT, to RDF, then thru my rdf->dot->SVG tools.
DanCon: found this article from the relax-ng TC home
jonb: nice article
danbri__: Step 2... "Specify mediators that allow interoperating between various services
danbri__: Step 3... "Describe the interfaces of wrapped services, e.g. using a state
danbri__: see the GINF pages for details.
danbri__: Step 3... "Describe the interfaces of wrapped services, e.g. using a state
danbri__: see the GINF pages for details.
danbri__: On "Coarse-Grained Communication
danbri__: ... "In the world of distributed objects, the idea was that you talked to an object through its methods, and objects are supposed to have this thing called encapsulation so that everything you can talk to about an object has a method."
danbri__: "Imagine if everyone went into the supermarket, bought a stick of butter, went home, came back, bought a quart of milk, went home, came back and bought a loaf of bread, and so on. If everybody did that, it would be a disaster."
danbri__: On why WSDL (Web Services Description Language) is incomplete: "Here's the problem: If these things are unpredictable, if the service that you sent a message to hasn't responded within two days, that's probably an error. Should they have responded within two minutes and they haven't, it's probably an error. You have to have some idea of the contract that's implicit in the service. Today, that's not part of WSDL or SOAP. It doesn't say, "
danbri__: ... the expected latency?" "What's the acceptable round-trip time before I come back to you?" Secondly, if I send you a message and you were going to send me one back when you're done, you have to tell the person who to send it back to. It's not clear they know how to send you back a message saying, I'm done, or I'm ready to go to the next step."
danbri__: "The tricky thing about WSDL today is it's just a set of messages. It's missing those two characteristics; it says what the messages are, but it doesn't say how you know what order to do them in?if anything happens to be known about the order.
danbri__: On building ODBC and other OLE controls at MS: "... One of the things we learned the hard way is you'd better write down the order in which things get done?until you write what we call state machines, this stuff just didn't interoperate worth a darn. "
danbri__: ... "In the world of distributed objects, the idea was that you talked to an object through its methods, and objects are supposed to have this thing called encapsulation so that everything you can talk to about an object has a method."
danbri__: "Imagine if everyone went into the supermarket, bought a stick of butter, went home, came back, bought a quart of milk, went home, came back and bought a loaf of bread, and so on. If everybody did that, it would be a disaster."
danbri__: On why WSDL (Web Services Description Language) is incomplete: "Here's the problem: If these things are unpredictable, if the service that you sent a message to hasn't responded within two days, that's probably an error. Should they have responded within two minutes and they haven't, it's probably an error. You have to have some idea of the contract that's implicit in the service. Today, that's not part of WSDL or SOAP. It doesn't say, "
danbri__: ... the expected latency?" "What's the acceptable round-trip time before I come back to you?" Secondly, if I send you a message and you were going to send me one back when you're done, you have to tell the person who to send it back to. It's not clear they know how to send you back a message saying, I'm done, or I'm ready to go to the next step."
danbri__: "The tricky thing about WSDL today is it's just a set of messages. It's missing those two characteristics; it says what the messages are, but it doesn't say how you know what order to do them in?if anything happens to be known about the order.
danbri__: On building ODBC and other OLE controls at MS: "... One of the things we learned the hard way is you'd better write down the order in which things get done?until you write what we call state machines, this stuff just didn't interoperate worth a darn. "
DanC_lap: Jelliffe and Borden discuss RDF syntax details.
DanC_lap: "Actually, the name of the rdf attribute is never written _n, because under
DanC_lap: XML production 41, a name can never start with a _ character."
DanC_lap: oops
DanC_lap: production 5 seems to allow <_1/>
jonb: that's correct in fact this test case shows that
DanC_lap: "Actually, the name of the rdf attribute is never written _n, because under
DanC_lap: XML production 41, a name can never start with a _ character."
DanC_lap: oops
DanC_lap: production 5 seems to allow <_1/>
jonb: that's correct in fact this test case shows that