On Monday, February 26, 2007, at 05:34PM, "Anne Thomas Manes" <[EMAIL PROTECTED]> wrote: >In HTTP, the self-description is the MIME type declared in the HTTP header. >Standard MIME types (i.e., registered types) describe the type of content >(image, text, XML, etc), but they typically don't provide semantic >information, such as what's in the image, what's in the text, or what's in >the XML.
Right, they are MIME types that are usuitable for the tasks we discuss here. It does not come as a surprise that more expressive MIME types are not the majority on the Web today as it has been mostly for human consumtion. The concept however is not affected by the lack of such mime types (UBL is a god example of something that should definitely have a MIME type). But there are already some examples such as application/atom+xml , application/atomsrv+xml or text/calendar. > >As I indicated in an earlier response to this message, I argue that the >qualified name of the root element in a SOAP Body is a much more descriptive >self-description than a MIME type. And why is that so? > >You can create your own MIME types, and in fact define a MIME type for an >OASIS UBL document or a private schema negotiated out of band. I haven't >looked it up, but I'd guess that EDI standard documents are, in fact, >registered MIME types. Yes, they are. > >The probelm with creating your own MIME types is that standard web clients >don't know how to process them. But the situation is not any different with XML vocabularies. One of the main reasons why the Web works is >that MIME types are standardized, and all web clients know what to do when >they get a JPEG in response to a GET. But what are they supposed to do with >a privately defined MIME type? The main issue here is that the Web *can* grow. A party can make up a mime type, I can look it up and implement software. If the format meets many peoples needs, it will grow and eventually become a standard or de-facto standard. An hey - let's all go and engage in crafting MIME types for the stuff we need for doing business with Web tecvhnologie. (A far more fruitful task, IMHO, than defining standard interfaces for Web Services that need to worry about all the behavior and invocation order unnecessities. Jan > >Anne > >On 2/26/07, Dennis Djenfer <[EMAIL PROTECTED]> wrote: >> >> Jan Algermissen wrote: >> >> >> On 25.02.2007, at 22:55, Eric Newcomer wrote: >> >> Jan, >> >> In that case I would expect WebSphere MQ to be compatible with the Web >> since it has a uniform interface. Correct? >> >> >> Messaging systems have a uniform interface[1], yes. I am not sure what >> you mean by 'compatible' though. Messaging systems are a different kind of >> architectural style than the Style of the Web is. Following one of REST's >> constraints doesn't make a style the REST style. The Pipe and Filter style ( >> e.g. Unix shells) for example employes a uniform interface, too. But that >> style is not the Web style either. >> >> Do massaging systems constrain the messages to be self descriptive (do >> they require the payload to be of standardized formats)? Message >> self-descriptiveness is also one of the very important one of REST's >> constraints. >> >> Hi Jan, >> >> It seems like you're saying that a message is self descriptive if it's >> standardized, right? >> >> I'm curiouse about your defintion of a self-descriptive message: >> Is an order message that follows a schema that has been defined by OASIS >> UBL self-descriptive? >> Is an order message that follows a schema that has been negotiated out-of >> band between many organizations in a specific business domain >> self-descriptive? >> Is an order message that follows a schema that has been negotiated out-of >> band between two organizations self-descriptive? >> Is an EDI message that follows a structure that has been defined by an EDI >> standard organization self-descriptive? >> >> >> // Dennis >> >> Jan >> >> >> [1] You can of course also abuse those to tunnel commands >> >> >> >> >> Eric >> ----- Original Message ---- >> From: Jan Algermissen <[EMAIL PROTECTED]> >> To: [email protected] >> Sent: Sunday, February 25, 2007 2:53:27 PM >> Subject: Re: [service-orientated-architecture] SOA Pizza Order Surprises >> >> >> Eric, >> >> your posting 'deserves' a more detailed reply, so sorry for only sending >> a short comment (I still have a pile of work on my desk for tonight). >> On 25.02.2007, at 18:23, Eric Newcomer wrote: >> >> It is just hard to believe that the lack of uniform interfaces in SOAP and >> WSDL is the cause of all the disconnect with the Web. >> >> >> The lack of a uniform interface (the plural doesn't really make sense >> here, does it?) is contrary to the architectural style of the Web. That is >> just an undebatable >> fact. An architecture that does not employ a uniform interface can never >> be of the REST style and an architecture that does not specifically >> constrain itself to >> HTTP's set of methods on all objects is necessarily disconnected from the >> Web. >> >> Jan >> >> >> (And, yes, GET /foo/lauchMissile is not HTTP's GET, it is tunneling the >> launchMissile invocation through GET) >> >> >> >> >> >> ------------------------------ >> Bored stiff? <http://us.rd.yahoo.com/evt=49935/*http://games.yahoo.com> >> Loosen >> up... >> Download and play hundreds of games for >> free<http://us.rd.yahoo.com/evt=49935/*http://games.yahoo.com> >> on Yahoo! Games. >> >> >> ------------------------------ >> >> No virus found in this incoming message. >> Checked by AVG Free Edition. >> Version: 7.5.446 / Virus Database: 268.18.4/702 - Release Date: 2007-02-25 >> 15:16 >> >> >> >> >
