Shibboleth Developers

[Karsten Huneycutt <>]

Hello --

On 03 Feb 2010, at 13:45 , Chad La Joie wrote:

> On 2/3/10 1:27 PM, Karsten Huneycutt wrote:
>> If the goal is ready access to data connector definitions for deployers, 
>> why not break those out into a separate file instead?
> 
> Well, you can already do this.

Then maybe that should be the default?

>> My concern with splitting these two definitions apart centers around 
>> adding new attributes to be resolved and having to add this in two 
>> separate yet very coupled files.  Plus, these two functions, while not 
>> necessarily the same in code, are logically part of the same attribute 
>> definition -- one won't function without the other.
> 
> I have a feeling this is going to be one of those things where any 
> suggested changes will make the file(s) make more sense to some and less 
> sense to others.

Add me to the "less sense" column.  To me (and also from a practical 
perspective of adding awareness of a new attribute), the attribute definition 
isn't complete without the encoding rules.   While I agree that the current 
file is a little verbose and hefty, I think moving the encoding definitions 
really only moves the verbosity and sacrifices clarity.

KH

-- 
Karsten Huneycutt
ITS Identity Management, UNC Chapel Hill