| It might be best to add a new name, rather than renaming it entirely.
This would be less likely to break other programs.
The name of an attribute is not very significant, it is mainly used by
our user interfaces (dxim and infobroker). Other applications dont know
how to read our schema files, so have to do their own OID to name
translations. It might be possible to use a different customised schema
for each application/user/system to get different translations for
different environments.
If you are using AltaVista Directory then you can't change the schema.
However V1.0A will accept the alternative name 'mail' for
rfc822mailbox, to allow working with Netscape and Internet mail.
Andrew
|
| re: .1 & .2
Thank you for the responses to my question. Unfortnately, I did not
understand the customer's question/problem.
Randy;
Let me restate my problem:
I built my X.500 using the attribute rfc822Mailbox to contain an
individuals "internet" email address.
It seems that with the rapid expansion of LDAP to access X.500, the
attribute "email" is being used quite widely for Internet addresses.
A number of LDAP browsers that I have seen depend on the email
attribute.
So, my choices to allow wide successful interaction of my X.500 with
LDAP browsers seem to be:
- change my X.500 to use an "email" attribute for internet addresses
- duplicate the internet address in both rfc822Mailbox and email (not
good)
- Somehow have the X.500 treat the attributes email and rfc822Mailbox
as equivalent in the sameway X.500 treats "cn" and commonName. So that
information is stored in a single place.
The third option is the one I am interested in having you respond to.
|
| The name of the attribute is just a name. An attribute can have
many names, and can be called different things by different clients.
The DSA does not use the name - it uses the OID which is a unique
tag which is associated with the attribute.
The name has to be translated to an OID by anything accessing the
directory. In the case of the digital clients (including infobroker)
this translation is done using a schema file. Our schema file
format allows multiple names for an attribute, using the LABEL
clause. This is how cn and commonname are dealt with.
However, only one of these names is used for the reverse translation
(either the first or last in the list). So, with infobroker I think
an application can ask for an attribute by any name, but will get
it returned using the 'standard' name (It might be clever enough
to remember what the name the client prefers, but I dont think so).
There is no reason why you could not use schema files with a
different ordering to the attribute names for different applications.
So your standard schema file could use the attribute name
'rfc822mailbox', while the schema file used by infobroker could use
'email' (or 'mail', which is what netscape and internet explorer
use).
You could even have different infobroker servers with different
name translation tables so that different applications can use
different attribute names (for the same data).
AltaVista directory remembers the name the client asked for, so
any client that asks for specific attributes will get back what
it expects. If a client asks for 'all attributes' then it gets
them back with the standard name.
Some clients ask for an attribute using several different names.
That way they will get the data no matter what name the server
knows it as.
Andrew
|