I agree that ReST (Representational State Transfer) is orthogonal; in
fact it is an architectural style for ROAs.
Actually, in my email I mentioned the comparison between the "REST
Web technologies" and SQL. I meant the most common ReST
implementation which uses HTTP verbs as a small set of operations
implementing the CRUD pattern; in this case the architectural base
Interaction: HTTP (POST -> create, GET -> retrieve, PUT -> update,
DELETE -> delete);
Representation: XHTML, other XML, PNG, ...
Sure, it's definitely worth discussing OGC-REST. Thanks very much for
putting this document together and stimulating an interesting
To respond to some of your main points.
A valuable case of REST approach is the SQL. In fact the SQL commands
implement the CRUD functionalities, the FROM-WHERE clauses expresses
unique resource IDs.
It is certainly true that SQL can be used to express CRUD functions.
However, I don't think it's true that every SQL command is a CRUD
operation because SQL can do lots of things that simple CRUD and REST
can't do (easily). Only very simple SELECT operations can really be
considered as a "CRUD read", for example. Most SELECT operations will
involve more queries that are more complex than simply getting a
resource. They will involve subselecting and aggregating different
resources to produce a unique result. This is what I think of as the
distinction between a "read" and a "query" (a read is where you are
getting a single, pre-existing resource - perhaps a single row in a
table in a database, for example).
Therefore, SQL can certainly be used in a REST implementation, but SQL
is much wider than REST. How would you express the equivalent of an
inner join in a RESTful URI syntax?
Our point is whether it is possible to use the REST Web technologies
for Geodata infrastructures as SQL technology has been used for DB
I don't think this is a fair comparison. SQL is the query language
for databases. REST is not a query language for geodata: it is an
architectural style for web services. The OWS protocols are
effectively query languages for geodata (which may, of course, be
implemented using SQL on the server side). Therefore the geodata
analogy for SQL is the OWS syntax. REST is an orthogonal concern, in