Re: [wsrp-interfaces] Custom user profile items

Our past experience has shown that a
common reason arguments in a particular area keep going round and round
the same material is that we have not properly scoped the question/issue
being addressed. It became clear on yesterday's call that this is happening
in this case. The hope is that if we can scope the issue, building a consensus
will become easier.

Rich 

Subbu Allamaraju <>

08/31/05 09:01 PM

To



cc

Subject

Re: [wsrp-interfaces] Custom user profile
items

I see this stalemate as a lack of consensus between
solutions. I suggest 

that we try to reach some consensus between the following:

a. Drop custom user profile items from the spec altogether.

b. Change UserProfile/UserContext types to reflect custom user profile

items as in Andre's or my proposal.

c. Let implementations use the extension description as in Rich's proposal.

d. Expect that implementations will carry custom user profile items as

extensions, i.e. leave things as they are.

I'm purposefully narrowing the topic to just user profile items, and 

leave the broader debate on extensions to implementations.

Subbu

Rich Thompson wrote:

> 

> In reflecting on this mornings call, I think Mike is right in that
we 

> have not well enough defined the problem we are working on and this
is 

> complicating discussing solutions. My summary of the problem statements

> I have heard is:

>  

>  Richard: The v1 metadata for custom user profiles items is insufficient

> to locate and understand the items (i.e. reasonably support mapping).

>  Subbu: The v1 spec left where custom user profile items are
placed 

> ambiguous.

>  Andre: The protocol should cleanly support user profile schemas
defined 

> outside the TC.

> 

> I think Subbu's problem statement is a narrowing of Richard's with
a 

> strong indication of a preferred solution. I think Andre's statement
is 

> the broadest of the three proposed problem statements. If we can settle

> on which one(s) of these, if any, we are willing to accept as problems,

> it should help scope the discussion/solution in a productive manner.

> 

> Rich