Re: registries, XML & EPP (again)

Hi all,

The prefix problem is actually a bug in our documentation:  no specific
prefixes are required.
Concerning the IETF namespace URI's :  we will use our own URI's for all
the schemas that we have modified.

Maarten Bosteels
software developer DNS Belgium

Hollenbeck, Scott wrote:

>>-----Original Message-----
>>From: [email protected] 
>>[mailto:[email protected]] On Behalf Of Klaus Malorny
>>Sent: Wednesday, October 05, 2005 6:26 AM
>>To: [email protected]
>>Subject: [ietf-provreg] registries, XML & EPP (again)
>>
>>
>>
>>Hi all,
>>
>>digging deeper into the EURid EPP specs I discovered 
>>something that is in my 
>>humble opinion a bit more serious than the prefix problem. It 
>>looks like that 
>>they took the XML schema files of the domain and contact 
>>objects (RFC 3731, 
>>3733) and deliberately modified them, like adding a new 
>>contact type or changing 
>>whether elements are mandatory or not. I am not completely 
>>aware of all changes, 
>>as the schema files are not available yet. While the changes 
>>themselves are not 
>>a problem at all, EURid still associates these new schemes 
>>with the IETF 
>>namespaces, namely urn:ietf:params:xml:ns:contact-1.0 and 
>>urn:ietf:params:xml:ns:domain-1.0. Isn't this something that 
>>should really be 
>>avoided for the sake of the clarity and compatibility? 
>>Technically, it is not a 
>>problem to use any other URI, like they did for their 
>>proprietary extensions anyway.
>>
>>Please don't misunderstand me, I do not want to start an 
>>EURid or whatever 
>>registry bashing here, but the least what we need is a 
>>standard that isn't one 
>>(for whatever reason). Lessons like SQL or HTML have shown 
>>where something can end.
>>    
>>
>
>More than just "should really be avoided".  "MUST NOT be done"!
>
>Those URIs are the unique, registered identifiers for the EPP schemas
>and namespaces.  They identify the specs documented in the RFCs.  Change
>the schemas and you have something new, different, and EPP-like that
>MUST NOT (in 2119 terms) be identified using the same URIs.
>
>-Scott-
>
>
>  
>