Full_Name: Howard Chu
Version: 2.3/HEAD
OS: 
URL: ftp://ftp.openldap.org/incoming/
Submission from: (NULL) (76.168.84.21)
Submitted by: hyc


Our current schema definition defines groupOfURLs as a structural objectclass.
This  matches the definition used by Netscape, where this schema originates.
However, the published documentation describes it as an auxiliary objectclass,
and indeed the Netscape/iPlanet/Sun/FDS/RHDS servers use it as if it were an
auxiliary.

http://www.redhat.com/docs/manuals/dir-server/schema/6.1/oc_dir.htm#1316814

This ambiguity arises because Netscape/iPlanet/Sun/FDS/RHDS don't actually
implement the LDAP/X.500 data model correctly. There are many areas where they
fail to implement the model, but in this specific case they simply don't enforce
schema definitions.

http://forum.java.sun.com/thread.jspa?threadID=5016864&messageID=9034636

And it appears that their developers didn't consider this a problem worthy of
attention.

But it's creating confusion for other users.

http://www.openldap.org/lists/openldap-software/200609/msg00121.html

I think the right thing to do here is change our definition to AUXILIARY since
that matches its actual usage. It won't match the published definitions, but
those definitions are demonstrably wrong.