Comparing the "Economics" of REST and RPC
Thu, 16 Aug 2001 12:01:48 -0400
Let's say there were a series of 'active' spiders. You'd go to Google and type
in a search. The engine would then kick off a spider run of likely candidates.
Spidered sites might themselves kick off spidering runs of other sites.
This is obviously contortions to enable a peery way of thinking about the web.
Imagine webtella: Alice sends "GET /webtella/hosts/metallica.mp3" to Bob, Bob
sends the same thing to Carol, Carol returns an HTTP response with an entity
body containing her IP:port, Bob returns that HTTP response to Alice. Alice
connects to that IP:port and issues "GET /files/metallica.mp3".
What I'm looking for are fundamental differences between these UI models, given
that the interface exposed to a spider can be considered a UI.
> The difference in interface complexity is an interesting comparison.
> You missed one variable, though -- time. REST has problems with broken
> resource identifiers, but that is comparable to lost objects. Typical
> RPC-like architectures also have to deal with the fragile base class
> and interface versioning problems -- issues that are hidden behind the
> REST interface because the client doesn't have to deal with strong typing.