Lists Home |
Date Index |
I've been musing on something lately, and wondered if
this has occured to anyone else...
There seems to be several types of 'query' interfaces at
the moment. i.e. a simple request/response interaction
with a server.
Type 1. Web based, non XML
The standard HTTP based querying that's been around for
years. Query parameters encoded as GET/POST, with response
in protocol specific format
Type 2. Web based, XML response
Again querying is based on HTTP, parameters encoded
as GET/POST, with response(s) in XML.
Type 3. Web based, XML request/response
XML document based query interface (e.g. XML-RPC,
SOAP over HTTP), and XML response
Type 4. Protocol neutral, XML request/response
The high end. Protocol agnostic. E.g SOAP.
Does anyone think this is a useful classification?
I'd also assumed that Type 3 would be the most
interoperable way to define a query interface. However
I'm not sure the assumption stands up: Type 2 actually
has a lower barrier to entry.
It's easier to repurpose existing software (CGI scripts) to
support this kind of interface, and involves no extra software
on the server side (i.e. no XML parser). It's also a useful
approach for providing interfaces that can be used by XSLT,
Type 2 can also be used to boot-strap into a purely
Type 3 environment.
Leigh Dodds, Research Group, Ingenta | "Pluralitas non est ponenda
http://weblogs.userland.com/eclectic | sine necessitate"
http://www.xml.com/pub/xmldeviant | -- William of Ockham