... Well said. Perhaps the problem is that industry is not fully cognizant of what it has been building. I m reminded that many programmers did not realizeMessage 1 of 16 , Jul 5, 2002View Source
> -----Original Message-----Well said. Perhaps the problem is that industry is not fully cognizant
> From: Paul Prescod [mailto:paul@...]
> There is nothing trivial about a the amount of money industry
> has spent implementing the REST architecture. I'm sorry if it
> seems that way to you. One of the main inventors of REST is
> also one of the lead designers of Apache (Fielding). So
> surprise, surprise, Apache has REST ideas deep down in its
> core. I see nothing trivial about pointing that out. PHP's
> developers are strongly influenced by the Apache group, so
> PHP has REST ideas deep down in its core too. etc. etc. for IIS.
of what it has been building. I'm reminded that many programmers did not
realize fully what they were doing with OO until the likes of Jim
Coplien and the GoF articulated the idea of Software Patterns.
Bill de hÓra
... True enough. We all learned web programming so incrementally. It was never approached as an architectural style, it was always tactical: gotta get thisMessage 1 of 16 , Jul 5, 2002View SourceBill de hÓra wrote:
>True enough. We all learned web programming so incrementally. It was
> Well said. Perhaps the problem is that industry is not fully cognizant
> of what it has been building. I'm reminded that many programmers did not
> realize fully what they were doing with OO until the likes of Jim
> Coplien and the GoF articulated the idea of Software Patterns.
never approached as an architectural style, it was always tactical:
"gotta get this information up today. Better give a different URI to
each thing so it can be bookmarked. Oops, using POST messes up the
'refresh' button. Better use GET." etc.
Come discuss XML and REST web services at:
Open Source Conference: July 22-26, 2002, conferences.oreillynet.com
Extreme Markup: Aug 4-9, 2002, www.extrememarkup.com/extreme/
... Sure, but that s what J2EE is according to Sun. A big beast. ... Yep, but the interop standard for EJB is RMI-IIOP, not RMI; historically that was inventedMessage 1 of 16 , Jul 5, 2002View Source
> -----Original Message-----Sure, but that's what J2EE is according to Sun. A big beast.
> From: S. Mike Dierken [mailto:mdierken@...]
> Sent: 05 July 2002 18:17
> To: firstname.lastname@example.org
> Subject: Re: [rest-discuss] Re: [xml-dev] ANN: Building Web
> Services the REST Way
> ----- Original Message -----
> From: "Bill de hÓra" <dehora@...>
> > J2EE is way (waayy) more complicated than SOAP;
> > EJB+Servlets+JDBC+JAAS+RMI-IIOP+JNDI+JMS...
> That's because it's doing different things.
> I wouldn't include IIOP (from Corba) or RMI (from Java) -Yep, but the interop standard for EJB is RMI-IIOP, not RMI; historically
> optional & usually available but not required in J2EE.
that was invented to keep the CORBA crowd happy. And it's essence is
synchronous, RPC and pass by value. I honestly don't know how well the
stubs and skeletons thing fits with REST, but otherwise RMI-IIOP doesn't
strike me as a RESTful protocol.
> JAAS is probably just asJAAS looks nice, it does much the same for authentication and
> core - I just don't happen to have any experience with it.
authorization that JNDI does for directories and JDBC does for
datasources. Clients can declare controls impendent of implementations.
I wanted to use it an upcoming project, but I didn't think there was
time for me to pick it up properly.
Bill de hÓra
... I agree with your position that any web development platform could be called a REST toolkit. However, I think a more RESTful toolkit should have typesMessage 1 of 16 , Jul 7, 2002View SourcePaul wrote:
> I don't see the problem. REST is the architecture of the Web and a RESTI agree with your position that any web development platform could be called a
> toolkit would be a piece of software that helps you build the mediating
> layer between your application and the Web. AFAIK, that's what (e.g.)
> J2EE and Zope are used for. Now that XML has come along it makes sense
> to want to extend these things. Just as they had native knowledge of
> HTML, you might want to give them native knowledge of XML and RDF. And
> maybe newer toolkits could better enforce best practices. But overall, I
> see no problem calling any web development platform a REST toolkit. REST
> is not something new.
REST toolkit. However, I think a more RESTful toolkit should have types defined
in its namespace that correspond to Fielding's documented approach if it's going
to make programming RESTfully feel more concretely similar (and IMO easier to
comprehend the REST style).
For example, here are some of the objects that are being implemented in my
toolkit (sorry I keep bringing up my own vaporware, but it's the most concrete
thing I have right now): CObject, CString, CList, CTable, CResource,
CRepresentation, CConnector, CStream, CRequest, CResponse, CComponent, CFilter,
Add a handful of sample apps to a class framework like this and you've got a
nice simple way to communicate the essence of the RESTful style.