mca blog

now that i've worked through a handful of issues including adding properly parsing nested resource URI (using regex), supporting create for PUT (etags for the lost update), and integrating GET/HEAD as internal calls, i'm starting to see a new level of abstraction for the framework.

so far, i've focused on a solid SqlXml and File handler set that makes reading/writing resources easy and reliable. this does not include any custom workflow - just the basic CRUD-type work. that's fine. but now i want to expand the framework make is easy to start building apps that have additional workflow activities on the server when needed (accept the POST, if it's valid, store it, send the user an email confirmation, and POST a message to the work queue for further processing, etc.).

i don't want to try to create some internal scripty crap. just make the existing class library easy to use when creating custom resource handlers. this means abstracting things like caching & 304 support out of the primary code pile (or at least minimizing their appearance for coders); making support for etags on PUT as transparent as possible (a kind of IfExists() that uses headers and GET/HEAD) and improving the pipelining of requests internally to make sure they are as robust as succinct as possible.

all good stuff. all doable.

once that is working, it should be even easier to create custom resource handlers that can get *real work* done on the web!