I'd like to provide a very lightweight method to describe my web services. No WSDL, WADL or similar complex stuff. Just a mapping from the web service's base URL to a simple description which can be detected and parsed automatically. Basically
- The name of the service
- A link to some human readable documentation (optional)
- Possibly a contact address
- Usage conditions, URL patterns etc. are less needed to keep it simple.
The best I could find is OpenSearch Description Documents, but first it could be simpler with plain text config files, and second there is no standard how to connect these service description documents to arbitrary web services (only autodiscovery in ATOM/RSS and HTML).
I thought about creating my own format and provide it similar to robots.txt
:
$ curl http://example.org/services.txt
service: http://example.org/foobar
name: Foobar API
about: http://example.org/foobar-doc.html
service: http://example.org/some/unapi
name: Example ORG's unAPI
# additional fields, to be discussed
standard: http://unapi.info
pattern: ?id={id}&format={format}
So if you know the base URL of a service, such as http://example.org/some/unapi
, you can look up at least the name in the service description file.
Instead of defining my own standard, however, I'd like to know whether there is already something similar to reuse.