[Date Prev][Date Next] [Chronological] [Thread] [Top]

subtypes RE: draft-smith-ldap-inetorgperson-02.txt



I have questions about how AttributeDescriptions are supposed to work.

rfc 2251 says they should be treated as subtypes of the attribute. How
subtypes work is described in X.501:

----- 
Where subordinate specialized types are selected to be returned as part of a
search result these types shall be returned if available. Where the more
general types are selected to be returned as part of a search result both
the general and the specialized types shall be returned, if available. An
attribute value shall always be returned as a value of its own attribute
type.

All of the attribute types in an attribute hierarchy are treated as distinct
and unrelated types for the purpose of administration of the entry and for
user modification of entry content.
-----

this next part is where the problems start:

-----
For an entry to contain a value of an attribute type belonging to an
attribute hierarchy, that type must be explicitly included either in the
definition of an object class to which the entry belongs, or because the DIT
content rule applicable to that entry permits it.
-----

X.500 will define subtypes with their own OID numbers. Ldap doesn't. Does an
attributeDescription with an option map to the same X.500 attribute (with
the same OID) or to a different attribute with a different OID? Operations
done on attributeDescriptions using X.500 protocols may not be able to
distinguish between attributeDescriptions. Perhaps X.500 ldap servers must
implement attributeDescriptions as Contexts. Or must an X.500 ldap server
insist that the ldap schema definition of attribute subtypes use an OID
which is different from the supertype's OID? Or is the intent that even in
ldap, attributes which are subtypes and attributeDescriptions must be
defined with their own OID?

rfc 2252 (section 4.2) implies that attributeDescriptions must be explicitly
defined in the schema. But 2251's description of the ";binary" option
implies that this value is automatically known to ldap servers (" .. servers
store any particular attribute in a single format .."). This means that for
the ";binary" option, attributes are not treated as subtypes but as one
single attribute type. Perhaps any attribute defined with syntax Binary will
be treated this way but those that are not will treat the attribute as a
regular attributeDescription (i.e. 2 separate values)? Also, if the
attribute is defined as Binary, only the ";binary" version of the
attributeDescription will be returned in a search operation but the
";binary" option is not necessary as parameters of the add, modify or search
operation. Is this correct?

Chris.

> ----------
> From: 	Jim Sermersheim[SMTP:JIMSE@novell.com]
> Sent: 	Friday, February 26, 1999 1:42 PM
> To: 	mcs@NETSCAPE.COM; Thomas.Salter@unisys.com
> Cc: 	ietf-asid@NETSCAPE.COM; ietf-ldapext@NETSCAPE.COM
> Subject: 	RE: draft-smith-ldap-inetorgperson-02.txt
> 
> >>> "Salter, Thomas A" <Thomas.Salter@unisys.com> 2/25/99 10:41:05 AM >>>
> 
> >I've always been confused about language tags.  Is first example showing
> two
> >different values of the attribute displayName, or two different
> attributes
> >which are subtypes of displayName ?  I read RFC2251 to say that they are
> >subtypes, in which case the singlevalued rule can apply to each. 
> 
> This is my understanding as well (from 2251: An AttributeDescription with
> one or more options is treated as a subtype of the attribute type without
> any options).  I'm having a hard time finding documentation that shows
> someAttr;optionA: JB Briggs
> someAttr;optionB: Joe Bob Briggs
> as a single attribute with two values.
> 
> If they are treated as subtypes of someAttr, what does it mean to delete
> the someAttr attribute from an object entry?  As far as I'm aware,
> deleting an attribute doesn't delete subtypes of that attribute as well.
> 
> Jim
> 
> 
>