Re: Issue 302: Graph edges that do not terminate

On Tuesday, Sep 10, 2002, at 16:06 US/Eastern, 
noah_mendelsohn@us.ibm.com wrote:

> I agree with everything you wrote regarding RPC and encodings that run
> into headers.
:-)

>   I was only responding to your original suggestion that in
> the case of RPC, there could be at most one root. My interpretation is:
> "yes" insofar as you do know where to start decoding the method name 
> and
> arguments, but "no" in the sense that a header could in principle
> introduce another sourceless edge into the node (I.e. typically the
> struct) that holds the arguments.  I agree this is more or less 
> harmless,
> but in this sense there can be more than one root, I think.
>
Agreed.

Marc.

--
Marc Hadley <marc.hadley@sun.com>
XML Technology Center, Sun Microsystems.