UWS as a REST protocol
dtody at nrao.edu
Mon Feb 26 13:19:07 PST 2007
Hi Doug -
I ran into the STREST stuff a while back, too. Basically what they
argue is that some folks use GET for what are really arbitrary function
calls, and they call this STREST. It is better than SOAP in some ways
in that at least one can use standard protocols, but it is not really
REST as the calls may change every time. I conclude that a GET can be
used either way.
As a counter example, if we have an access reference such as
Then I suggest this is probably about as RESTful as most URLs which
reference static files (so long as the dataset identifier is persistent).
The main point of REST appears to be simplicity, elegance, and reuse of
standard protocols and infrastructure.
On Mon, 26 Feb 2007, Doug Burke wrote:
> On Feb 26, 2007, at 2:32 PM, Doug Tody wrote:
> > Hi All -
> > This sounds very promising, I'm all for it (but you knew I was going to
> > say that right?).
> > On Mon, 26 Feb 2007, Matthew Graham wrote:
> >> If we are going down this path then I want to strongly reiterate (to
> >> this group and other IVOA groups) that we do it properly, i.e. pure
> >> REST
> >> interfaces and not the half-baked "well, it uses HTTP GET and verbs"
> >> that a lot of people mean by REST.
> > Actually, it is not all that easy to determine what REST really means.
> > I think it is REST so long as GET URLs are reasonably persistent, are
> > cacheable, and do not cause side-effects. This does not necessarily
> > mean that we cannot have GET URLs which contain parameters (e.g., a DAL
> > access reference). I have no problem with using POST for anything
> > which
> > causes side-affects, or for passing data elements.
> > - Doug
> Well, I've just been reading about STREST , to add yet more
> acronym-induced confusion to the whole "what is REST" discussion 
> (and to whether it actually is different to REST ).
> A different Doug
More information about the grid