Re: modeling problem
Re: modeling problem
- Subject: Re: modeling problem
- From: Chuck Hill <email@hidden>
- Date: Tue, 31 Mar 2009 15:37:01 -0700
On Mar 31, 2009, at 2:33 PM, TW wrote:
On Mar 31, 2009, at 10:06 AM, Chuck Hill wrote:
On Mar 30, 2009, at 4:29 PM, TW wrote:
My goal was to have an entity/class (where InetOrgPerson is
super) in a "middle" framework that can be used to wrap a
relationship between ldap user and database employee (each
modeled in their own frameworks). So, the way I have it set up
there's really no difference between InetOrgPerson and
GenericUser except that GenericUser has the relationship to
<database employee> and InetOrgPerson does not.
I still wonder if GenericUser should have a relationship to
both InetOrgPerson and DatabaseEmployee.
With that approach, it sounds like GenericUser would have to
have it's own data table separate from the other two?
Doesn't it need to anyway for the relationship to DatabaseEmployee?
Not really. Time for ascii sheep. This is what I have.
LDAPModel ---to build path--> UserModel <--to build path----
DBModel
======== ========= ================
InetOrgPerson GenericUser <-------------> Employee
(InetOrgPerson/parent)
The only relationship (on the right) uses "employee_id."
That is in the LDAP information?
Yes, InetOrgPerson has "employeeNumber." We use FileMaker-based
tools I created for adding accounts to OD. They call DSImport to add
records to OD as needed and they're set up to push the employee_id
into that attribute.
Actually, there's a scenario right there that supports your
suggestion to have an intermediary GenericUser table since that is
essentially the same position the old FileMaker database holds -
collection of "everyone."
Worth a further thought.
This approach makes the ldap side of things the starting point
since that is the store that knows what type of user the person
is. If ldap authentication is successful, I query for the matching
GenericUser which is essentially an ldap person since GenericUser
inherits from InetOrgPerson.
That sounds like a correct use of inheritance. I think you are
going to need a restricting qualifier on InetOrgPerson that is
always false.
I tried the (1=0) restricting qualifier and EOGenerator did not like
it and coughed up an error about being unable to parse the
qualifier. Good thought though.
you could try some other qualifier based on a attribute that will
always be false.
Well, I've got a version in my repository the way it is. I'm trying
to model it the other way around now and we'll see how that works.
Let us know!
Chuck
Then, I can test for the related employee record. If there, foo,
if not, bar - depends on the app.
It sounds like your preferred approach would be to create another
table for GenericUser making it a database entity instead of
inheriting from InetOrgPerson. That could work too. The problem
there is then I have two entities, InetOrgPerson and GenericUser
that have to have similar attributes, including "usertype,"
"email," etc. And, when a new Employee is created, I'd have to
force the creation of a GenericUser.
It sounds like you are on the right track. I think.
I can see advantages each way and it's too bad I'm somewhat at the
mercy of the greater campus.
Hence the lure of herding goats and sheep, living on wine and cheese.
Chuck
--
Chuck Hill Senior Consultant / VP Development
Practical WebObjects - for developers who want to increase their
overall knowledge of WebObjects or who are trying to solve specific
problems.
http://www.global-village.net/products/practical_webobjects
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Webobjects-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden