I just finished reading RESTful Web Services. It is an amazing book and I think it will play a very important role in defining main principles of the next generation of the Web. The authors of the book introduce the Resource-Oriented Architecture (ROA) as an architecture for building the resource-centric programmable Web. “Resource” is a fundamental concept in this architecture.
“A resource is anything that’s important enough to be referenced as thing in itself… What makes a resource a resource? It has to have at least one URI. The URI is the name and address of the resource…”
“… A resource can be anything a client might want to link to: a work of art, a piece of the information, a physical object, a concept, or a grouping of references to other resources… The client cannot access resources directly. A [ROA-based] web service serves representations of a resource: documents in a specific data formats that contain information about the resource…”
ROA defines principles of organizing data sets as resources, approaches to designing representations of these resources and main operations on these representations.
The key concept of the Subject-centric computing (SCC) is a “Subject” which is defined as “anything whatsoever, regardless of whether it exists or has any other specific characteristics, about which anything whatsoever may be asserted by any means whatsoever”. This definition is very close to the definition of a “Resource” in ROA.
But there are important differences between ROA and SCC main goals. The Subject-centric computing is less concerned with managing resource/subject representations and using universal HTTP operations such as GET, POST, PUT and DELETE to manipulate resources. SCC assumes that there are a lot of different data sets/documents (at least potentially) which describe or reference the same subject. With SCC, our main concern is in identifying subjects reliably and in bringing together different pieces of information related to the same subject.
As with ROA, we use (resolvable) URIs to identify Resources/Subjects. But in the case of SCC, we promote usage of Published Subject Identifiers (PSIs). If we have a subject that is not a “digital information item”, its PSI should be resolvable to a special kind of a “document” – Published Subject Descriptor (PSD). Each PSD provides a human readable description of a subject which is enough for distinguishing this subject from other subjects. Using ROA terminology, PSD is a special kind of a representation that is introduced to convey “identification” information about a subject.
Many other “documents” and data sets which contain various assertions about the same subject can exist on the Web. SCC is concerned with providing ability to collect these various assertions into the 360Â° view of the subject. PSIs are one of the main mechanisms to achieve this goal.
With SCC, we do not have a luxury of doing point-to-point data integration each time when we have a new data set. That’s why we rely on universal representation formalism which is an important part of ISO Topic Maps standard. Topic Maps provide also a universal merging mechanism that takes care of integration of various data sets published using an interchange syntax such as XTM.
One of the main goals of SCC is to support “associative nature” of human thinking. ROA is satisfied quite often with “shallow” representations of associations (with the “a” HTML tag, for example). SCC is more targeted to semantically rich representations of relationships between subjects. Topic Maps help to represent and manage such relationships as “instance-type”, “supertype-subtype” and thousands of domain-specific association types. Representations of these relationships are available for processing at the semantic level. It makes possible to implement integration scenarios which are “unthinkable” with HTML-like representations.
But in general, ROA and SCC are complementary architectures and can be successfully used together to build exciting applications and environments