[
Lists Home |
Date Index |
Thread Index
]
On Apr 5, 2005 4:29 PM, Jan Algermissen <jalgermissen@topicmapping.com> wrote:
>
> With the single-service approach I have something like this:
>
> POST /myService
>
> person-id: 778
> method-name: getAddress
>
> ...service returns address....
>
> With REST you'd have:
> GET /persons/778
What if I'm looking for John Smith and don't know his person-id? Or
whether it is Smythe or Smyth... or Jon, or Jan ... or which office
he works in? A centralized name disambiguation and forwarding
service would come in very handy.
Or, consider those acm.org forwarding addresses many of us use ... why
not expose the actual email address? I suspect a) we don't want to
expose the real address to slimeball spammers, b) we want to be able
to keep the same logical address if one changes "implementation
details" such as employer, ISP, etc. c) we want a centralized point of
management (at acm.org) for this and other personal metadata we choose
to make public. Similar reasons apply to the address example.
> HTTP semantics define GET responses as cacheable, so any generic HTTP
> intermediary knows it can cache the response.
>
> That improves scalability, IMHO.
*if* your users do far more read operations than create/update/delete
operations, and there is a high probability that a document that is
retrieved once will be requested again soon.. Very true on the Web.
Not true for data enty applications. Nightmarishly difficult in the
distributed read/write case.
|
