Re: requirements for accessing schema in DIT

["Kurt D. Zeilenga" <Kurt@OpenLDAP.org>]

At 04:47 PM 2/6/2006, Howard Chu wrote:
>Yes.... In practice 

I'll note that I see slapd(8) eventually moving to a split
schema model where certain items (e.g., syntaxes, matching rules)
are managed server wide (in say "cn=Server Schema" and other items
(e.g., DIT content rules, DIT structural rules, name forms) are
managed on a subtree basis (possible restricted to a database instance)
(in say "cn=Subschema,DatabaseSuffix", with some items (e.g., object
classes, attribute types) managed either way.

Of course, we should still support the X.500/LDAP subschemaSubentry
discovery model.  Basically, when viewing (but not updating)
"cn=Subschema,DatabaseSuffix", items from "cn=Server Schema"
would be transparently included.

I see two factors driving slapd(8) in this direction:
  - complete replication (user application data + meta-data,
        including schema)
  - advanced X.500 features (e.g., DIT structural rules, etc.)

Kurt