Lists Home |
Date Index |
On Tuesday 12 February 2002 06:27 pm, Michael Brennan wrote:
> > This is because REST is an architecture, not a protocol.
> Yes. I misspoke, there. But aside from a few misplaced words, I
> think my point is valid. The point I was making is that REST tells
> us the protocol should be quite visible to the application
> architecture (and vice versa), whereas typical patterns used today
> by web developers is to minimize that intrusion and hide the
> protocol from the application level design (and vice versa).
I'm not sure that REST says that, though Roy's thesis (from memory)
*does* say that choice of protocol should be defined by choice of
networking architecture, which should be defined by choice of
application architecture etc. I think this is sensible.
> There are other issues with REST, here, but I don't see complexity
> as one of them.
REST is not complex (especially in the most abstract as a set of
design constraints). REST over HTTP *is* complex.