Re: Why MD5 Headers are Imperative

Yes, and when I was diagnosed with Osteoporosis and given a prescription for 
Fosamax, it wasn't until after I encountered the side-effects that I 
discovered that it was initially approved for women based on testing of a 
sample of 9,600 post-menopausal women.  Two years later, I assume after 
someone at the FDA accepted a bribe, it was approved for men.  But nobody 
ever tested it on us, even though we're 20% of Osteoporosis patients.

Your sample group was perhaps restricted to implementations which would never 
encounter the same problem as me, that's hardly evidence that my example is 
invalid or that my code works for all Atom implementations, not just the 
ones where it's assumed the editor is also the publisher.

-EJB

>-----Original Message-----
>From: James M Snell [mailto:jasnell-Re5JQEeQqe8AvxtiuMwx3w@xxxxxxxxxxxxxxxx]
>Sent: Thursday, August 17, 2006 06:29 PM
>To: atom-protocol-O6w3ZxSwtmQ@xxxxxxxxxxxxxxxx
>Subject: Re: Why MD5 Headers are Imperative
>
>
>I've implemented no less than four APP server implementations and three
>clients.  I've tested my clients against about seven different server
>implementations, including three implementations from other vendors. I
>have yet to encounter any problem with the text you are claiming is
>broken.  Working code counts for a lot.
>
>- James
>
>Eric J. Bowman wrote:
>> I am astonished at the lack of concern here for the key element of what I'm 
>> saying:
>> 
>> I have a fully RFC-compliant Atom deployment.  It only breaks when I attempt 
>> to implement _your_ spec.  With _my_ spec it works beautifully and is still 
>> a fully RFC-compliant Atom deployment, with PUT.
>> 
>> Doesn't working code count for anything?  Not that I can back that up 
>> publicly just yet, but how about some benefit of the doubt?
>> 
>> -EJB
>> 
>> 
>> 
>> 
>
>