16074Re: [rest-discuss] RESTful Reference Architectures?
- Aug 1, 2010Bryan Taylor wrote:
>You may have a point there. REST has no "reference architecture", it's
> I suspect that the difference in viewpoints are the reason that
> people versed in REST and people versed in SOA seem so often to talk
> passed each other. REST as an architectural style is simply not
> expressed in terms remotely close to those that describe the service
> ecosystem viewpoint and IT ownership viewpoint that OASIS used to
> express a SOA reference architecture.
a style, whose reference instantiation is the Web, but is not limited
to the Web. The Web is an enabler of a variety of system architectures.
Each system is an instantiation of some architectural style, most of
which are undefined.
IEEE 1471 is a means for describing such architectures. The RA for SOA
is a good example. But I don't think it applies to defining anything
but instantiations of the Web or of REST (or of some other style). SOA
is an instantiation of the Web, but not of REST.
My demo website is in no way meant as a reference architecture for REST.
When all is said and done, I may very well use IEEE 1471 to describe it
as a reference architecture for CMS/wiki/weblog/forum operations and
integration, as a RESTful alternative to CMIS.
Such a reference architecture would describe an instantiation of both
REST and Web styles. But it would only be one possible REST or Web
instantiation of CMS/wiki/weblog/forum operations and integration, not
the be-all and end-all.
That's all IEEE 1471 can do -- describe reference architectures which
may be instantiations of REST, Web or other styles. I don't believe it
capable of describing architectural styles, i.e. as a replacement for
the WebArch document Jan linked to, or Roy's thesis.
- << Previous post in topic