Context "by value" versus "by reference"

["Jim Webber" <Jim.Webber@newcastle.ac.uk>]

Hi guys,

As regards the mechanics of "by reference" contexts (and deliberating
sidestepping the debate as to its perceived value), at the moment the
access mechanism for "dereferencing" a context is underspecified.

Would be it be possible to define a simple message exchange with the
host context service along the lines of:

<get-context>
  <context-id>a:b:c:d</context-id>
</get-context>

Which presumably would elicit one of the responses:

<context>
  ...
</context>

Or

<no-such-context-fault>
 ...
</no-such-context-fault>

Or an out-of-band fault (eg not authorised) as appropriate. 

This suggestion addresses access at the message level, rather than the
current scheme which seems to imply following a URI (usually a URL).

Jim
--
http://jim.webber.name