Re: [xdi] Implications of multiplicity on messaging syntax

[Drummond Reed <drummond.reed@xdi.org>]

On Sat, Sep 1, 2012 at 2:30 AM, Markus Sabadello <markus.sabadello@xdi.org> wrote:

My understanding is that if we adopt this proposal:
https://wiki.oasis-open.org/xdi/XdiMultiplicity

Then (independently of the small change I am proposing in the other thread), I think this means the correct messaging syntax would be:

@!2222$msg$(!1234)$do/$add/...

rather than

@!2222$($msg)$(!1234)$do/$add/...

(This is currently used in the latest XDI Graph Model/Patterns/Statements PDFs)

Markus

 

Exactly right. $msg would be a collection of messages, $(!1234) would be an entity singleton member of this collection, and $do would be a collection inside this entity singleton.

Note that multiplicity applies to link contracts as well, i.e., $do by itself is a link contract collection. So this causes me to wonder whether, if there's just a single link contract reference, it shouldn't be $($do), i.e.:

    @!2222$msg$(!1234)$($do)/$add/...

This would leave open the possibility that a single message could reference multiple link contracts for multiple operations. For example:

   @!2222$msg$(!1234)$do$(!1)/$add/...

   @!2222$msg$(!1234)$do$(!2)/$mod/...

   @!2222$msg$(!1234)$do$(!3)/$del/...

To associate each operation with the precise link contract that authorizes it, the link contract references would also use multiplicity, i.e.:

   @!2222$msg$(!1234)/$do$(!1)/{link contract #1}

   @!2222$msg$(!1234)/$do$(!2)/{link contract #2)

   @!2222$msg$(!1234)/$do$(!3)/{link contract #3}

What do you think?

BTW, it seems this is also closely related to the question about ordering of operations within a message that carries multiple operatons.

=Drummond