[INFOD-WG] More on ExecutionMode for DROP

[Fisher, SM (Steve)]

> -----Original Message-----
> From: infod-wg-bounces at ogf.org 
> [mailto:infod-wg-bounces at ogf.org] On Behalf Of Dieter Gawlick
> Sent: 17 July 2007 01:38
> To: INFOD
> Subject: [INFOD-WG] More on ExecutionMode for DROP
> 
> The reference implementation shows that we have to clarify 
> the ExecutionMode in DROP.:
> 
> Lets us assume that the request is to DROP an entry we some 
> property instances referencing this entry. We should drop 
> that entry and all the property instances referencing it in 
> all cases - otherwise the drop logic is somewhat strange. 
> 
> "DISABLE NEW' would ignore this entry, it is in the solution 
> set of the newly defined constraint; it remains in the 
> solution set of existing constraints. 
> 
> "CASCADE" is not needed any longer, we just remove the entry.
> 
> This makes the implementation much easier
> 
> Dieter

I still find this whole argument totally unconvincing. Even if we want
to we cannot make such a change to the spec as it is not just a
clarification but a real change and would require a new document.

DISABLE NEW simple marks the record in the database as disabled and it
is then no longer available to anybody to add to reference from a new
database record - however anybody with a reference to it via the EPR
can continue to use it. We should of course be careful in the
implementation with any tempoarray implelemntaion tables that they do
not attempt to violate this

Steve