[Cosmo-dev] cosmo 0.6.1 EIMML changes

[Randy Letness]

Jeffrey Harris wrote:
> It's a bummer that we miscommunicated about all this, I've implemented
> one thing in Chandler, you've implemented a different thing in Cosmo, we
> both thought we were working from the same page.  I'm not sure which
> side will have an easier time shifting, but I can certainly work on
> changing Chandler's implementation.
>
> I can change Chandler's modification translators to whatever works best
> for Cosmo, as long as somehow, any record associated with a modification
> includes an indication that it's about a modification and what that
> modification's recurrence-id is.
>   

I don't think we're that far off.  Cosmo still has to represent a 
modification as a separate note item.  We just need to decide on the 
protocol.  I'm fine with defining a noteModification and 
eventModification  record  type that each have two key fields 
(parentUuid and recurrenceId).  Cosmo will assign a random uuid to 
modification items, but will use parentUuid and recurrenceId to to 
lookup existing items.

The only problem I see is that in eimml, a recordset is related to a 
cosmo item by uuid:

<eim:recordset uuid="item1">
  <event:record xmlns:event="http://osafoundation.org/eim/event/0">
  ...
   </event:record>
...
</eim:recordset>


So for modifications, would the modification records be included in the 
parent item's recordset, or in its own recordset?  Since we aren't 
exposing the uuid, it seems like they would have to be included in the 
master's recordset.

-Randy