[security-services] Why use name-value pairs for modeling attributes?

Title: Credentials collection proposal

Sometime ago 
(Aug 26), Joseph Reagle had asked me the question (the text is 
mine):

 

----- Why do 
SAML attribute statements include an AttributeType which models 
the            

---- 
relationship between an namespace-qualified attribute name and XML value as a 
triple? 

---- Why not 
express this information as an XML fragment? The latter is the more standard 
------ treatment anyway.

 

 

In other 
words, instead of:

 

<Attribute>

  
<AttributeDesignator AttributeNamespace="http://www.finance.org/V1" AttributeName="CreditRating"/>

  
<AttributeValue>Good</AttributeValue>

</Attribute>

 

use:

 

http://www.finance.org/V1:CreditRatingGood</CreditRating>  

 

I recall 
considerable discussion around this topic around F2F#3 (BTW, are the minutes 
still available someplace in the Oasis web site?). My recollection is that 
we chose this treatment primarily to unify the syntax of attributes 
and attribute queries. This was based on a consensus to avoid use of a 
general XML query language.

 

The corresponding attribute query has the 
form:

 

<AttributeQuery>

   
<AttributeDesignator AttributeNamespace="http://www.finance.org/V1" AttributeName="Credit"/>

</AttributeQuery>

   

which has 
fairly obvious syntax and semantics. 

 

Maybe other 
SAML-ers other aspects of this discussion. Retrieving the minutes would 
also be helpful. 

 

- prateek 
mishra