RE: [provision] Attributes Correspondence--now with RACF.

["Raepple, Martin" <martin.raepple@sap.com>]

see my comments below
Martin

>-----Original Message-----
>From: Gary.P.Cole [mailto:Gary.P.Cole@Sun.COM] 
>Sent: Dienstag, 7. März 2006 15:57
>To: PSTC
>Subject: [provision] Attributes Correspondence--now with RACF.
>
>Still haven't added Liberty Person or Liberty Employee--Jeff Bohren 

Please find a list of all Liberty SIS PP / EP attributes in the
Excel sheet attached to this mail. I created this list in order to compare
it with our schema attributes. My plan is to submit our schema attributes 
by then end of week 13.

>suggested that we do that.  However, patterns are starting to form--or 
>maybe I've just been looking at this too long. ;-)
>It would be nice at some point to put these schema attributes and 
>correspondences into a database so that we could analyze data, report 
>results and people could contribute additional schemas or revised 
>correspondences.
>
>I'm trying to remove the emotion (as much as possible) from 
>the process 
>of selecting "standard" attributes. I have no problem special-casing a 
>few favorites (and the attributes that model standard capabilities 
>probably get a pass as "operational" attributes), but I'd like the 
>criteria for selecting the rest of the attributes to be as 
>objective as 
>possible.
>
>My biggest question is this:  In general, at what threshold of 
>functional correspondence (commonality) we should nominate/consider an 
>attribute to be "standard"?
>A) Every schema must have an attribute that functionally corresponds.
>B) The majority of schemas must have an attribute that functionally 
>corresponds.
>C) More than one schema must have an attribute that functionally 
>corresponds.
>D) One schema must have an attribute that functionally corresponds.
>
>What do you all think?
>

I would vote for option B) since A) might lead us to a very narrow and
C) and D) to a too broad list of attributes. Nevertheless, I am sure there will be
attributes where it's difficult to decide whether they 'functionally correspond' 
or not and where we have to decide on a case-by-case basis.

LibertySchemas.xls