new level of abstrction for exyus

2007-08-17 @ 01:46#

i've reached a new level of abstraction for the exyus RESTful framework. to this point, i've worked with single-depth-level resource definitions for data handling: /category/{id}, /weblog/{id}, etc. but now, i need to move to the next stage - added depth in the data-modeling of URIs: /weblog/{wid}/comments/{cid} for example. it's simple on the surface, but i want to make it as easy as possible to define nested resources like this. that means minimal coding, and the proper level of abstraction in the code.

i have a number of things to deal with here:

  1. matching the URI pattern to the proper class instance
  2. parsing the URI within the code segments for proper validation
  3. including the properly parsed URI segments in the execution of the HTTP method code
  4. handling the URI segments in constructing the cache-clearing details of the code

some of this is pretty straight-forward. the URI pattern matching is basic. providing the URI segments to the XML and XSLT engines is also pretty easy. i already have a tokenizer service that can parse placeholders in the XML and provide proper params to the XSLT engine. but handling them properly within the code execution and the cache-clearing portions of the code is a bit trickier.

so far, i've 'hard-coded' the resource handlers to work with a single {id}. but now i need two. and will quite possibly need three or four in the future. what i need is code that can handle a param array. i probably also need to look into named arguments for regexp. C# has this already and i just need to brush up on their use.

so there's some pretty important stuff here. when i get this done, i'll have a powerful resource-mapping capability built into the exyus framework. and that will be sweet!