On Tue, Aug 10, 2010 at 09:20:51AM -0500, Tim Donohue wrote:
> On 8/10/2010 8:27 AM, Bojan Suzic wrote:
> > In regards to HTTP Authentication, I have put it in the wiki and will
> > investigate further how it can be done.
> > For the other authentication methods, I am not sure which ones may be of
> > interest?
> 
> I think you should just concentrate on HTTP Auth to begin with.  Mark 
> Wood suggested adding other authentication options (using the DSpace 
> stackable auth) later on -- but, I think this could come later 
> (especially since you are running short on time in this GSoC period).

I don't mean to push on this issue -- I agree that we can release with
just HTTP Basic and integrate the rest later.  (We should have a JIRA
improvement issue for this, when the feature goes in.)  But I did want
to say that, unless a feature has rather unusual requirements, it
shouldn't have to answer questions such as "which authentication
mechanisms are of interest?"  DSpace provides infrastructure for
authentication and there should be a way to pass this task through
each UI to the core, to be handled by common code.

This is, IMHO, the proper way to think about constructing features for
a sizable product: look around aggressively to see which routine tasks
(such as AAA) have already been addressed, and use that infrastructure
zealously.  Users get consistent behavior across the product, and you
do less work which is unrelated to the task you set out to accomplish.

Authentication, for example, is tricky, but almost always BORING, and
yet it must usually be done and always done well.  It should be a
prime candidate for "passing the buck" to the core of the product,
where a common implementation can take care of the fiddly bits once
for all.

-- 
Mark H. Wood, Lead System Programmer   mw...@iupui.edu
Balance your desire for bells and whistles with the reality that only a 
little more than 2 percent of world population has broadband.
        -- Ledford and Tyler, _Google Analytics 2.0_

Attachment: pgpx9qEK86mqr.pgp
Description: PGP signature

------------------------------------------------------------------------------
This SF.net email is sponsored by 

Make an app they can't live without
Enter the BlackBerry Developer Challenge
http://p.sf.net/sfu/RIM-dev2dev 
_______________________________________________
Dspace-devel mailing list
Dspace-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/dspace-devel

Reply via email to