Mark has a nice post on REST issues.
Mark gives me too much credit for the lo-rest/hi-rest terminology - I saw it in a Nelson Minnar/Google deck and assumed that it was implicitly “not evil.” Little did I know...
More to the point, I completely agree with his assessment of the “real” issues at play.
In particular, the authentication/identity story is very tough (MD5-ing by hand is for the birds (thanks for nothing Flickr)).
Worse is the contentType/format issues - in SOAP we took XML as the “top” of our data model and inherited XSD (and sufferred mightily for it).
In HTTP/REST, MIME is the “top” of the data model and you wind up inducing XML (and JSON and RDF and...) if you need to build general purpose software. Ouch! Hence the lack of great and diverse tooling support - it's hard!
As for the “lo-rest/hi-rest” dichotomy that Mark has saddled me with, I do think the REST-afarians are missing an opportunity by not driving home the secret sauce that is HTTP GET.
It's a no-brainer and I sometimes think the REST folks are “breaking into jail” by dragging in the rest of Roy's dissertation.
GET is one of the most optimized pieces of distributed systems plumbing in the world.
It's an absolute/objective slam dunk.
No arguing/evangelism needed IMO.
GET is the classic “the first bag is free” kind of feature a platform builder dreams about.
Posted
Mar 07 2007, 03:20 AM
by
don-box