Re: KDC model and atomicity

On Mon, Jun 11, 2012 at 4:08 PM, Sam Hartman <[email protected]> wrote:
>    Nico> The failed authentication count attribute.  I suppose we could say
>    Nico> that it's a virtual attribute whose value may be computed from other
>    Nico> sources, then there'd be no atomicity implication.
>
> Hmm.  I thought the difference between a schema and data model implied
> this.  If there is ambiguity I'd prefer to clarify that so long as the
> data is representend in the schema, how it is done is up to the
> schema. I think even calling out issues like avoiding atomicity would be
> reasonable. But I'd prefer for it to be global not just focused on this
> one issue.

Well, LDAP's meta-schema is limited to the needs of LDAP as an access
protocol.  It doesn't address all sorts of things, such as computed
attributes, referential integrity, triggers, and so on.  This document
isn't specific to LDAP, I know, but LDAP might be foremost in readers'
minds.

Perhaps we can just add a minor amount of text saying something to the
effect that some/all schema elements from the model may well be
computed and/or mapped not only to another schema's elements, but also
to operations.

Nico
--
_______________________________________________
ietf-krb-wg mailing list
[email protected]
https://lists.anl.gov/mailman/listinfo/ietf-krb-wg