Future of SabreAMF
I've been asked a bunch of times what's gonna happen with SabreAMF.
The thing is, the goals of SabreAMF were a lean AMF-parsing library without any pre-cooked frameworks like AMFPHP or WebOrb. What I mean with that is that both of those products force you to structure your service classes in a specific way, whereas I needed to fit an Flash Remoting library in existing business logic for webservice.
I feel like I have succeeded in that goal. It has became what I wanted it to be, it's in use at a few high-traffic sites and no bug reports have came in in the last 4 months.
So what's next?
The next step is to put the version number on 1.0 and freeze the codebase, unless new bug reports come in. In the meanwhile I have been working on a more complete web service framework, that will make use of SabreAMF, and will actually looks more like AMFPHP's approach. I have to work out copyright issues before I can promise it will get open sourced, but I have a good feeling about it.
- Service-classes are specified using an xml file.
- Services are accessible using AMF, JSON, XML-RPC, SOAP, REST and PHP-RPC.
- VO-Classmapping configuration for Actionscript 2, Actionscript 3, SOAP and PHP-RPC.
- Automatic self-documentation using Reflection and PHP docblocks.
- Proxy-service classes, allowing you to proxy external webservices. (this allows you to access SOAP services using PHP-RPC, XML-RPC services using AMF, etc.)
If anyone thinks this is not a bad idea, and willing to contribute drop me a line, or post a comment.