Re: Collect Proposed wordings (Was: Can everyone be happy?)

At 9:17 AM -0700 6/22/00, Henrik Frystyk Nielsen wrote:
>  > This is a typical misuse of terminology by the few left on this list
>>  who do not understand the model in the URI specification.
>>  If a "namespace name" does not "identify" a namespace then
>>  how are these words being used?  Is a namespace nothing,
>>  because it is abstract?  Is there a complete inability here to
>comprehend
>>  something whcih is not a string of characaters?
>
>I think it should be noted, that the examples that David has put forward
>are not solved at all by fixed base. He wants to be able to say that
>http://WWW.W3.ORG is semantically different from http://www.w3.org even
>though this breaks in a number of ways and falls in the category of "don't
>do that". In fact, he has pointed out that he would prefer java class
>names instead.

Larry Masinter has contradicted you on exactly this point (I don't 
have a record as to who he was responding to). Argument from 
authority is probably not admissible, but given his evidence, and his 
authority, I think we can lay the case insensitivity issue in DNS 
names to rest permanently.

That's just not part of URIs.

At 10:41 PM -0700 6/20/00, Larry Masinter wrote:
>
>>  There is nothing HTTP specific in the description of how to compare URIs -
>>  it just happens that HTTP URIs "use" all features of the URI spec. This is
>>  why the section is called "URI Comparison". The rules are general for any
>>  URI.
>
>The description of URL comparison in the HTTP document was only for the
>purpose of describing the equivalence of URLs used in the HTTP protocol.
>It certainly wasn't intended to have greater scope of applicability and
>shouldn't be taken out of context as some evidence about how namespace
>names should be compared.
>
>To support this, I point to RFC 2557 which also passed IETF review as
>Proposed Standard, which uses byte-for-byte equivalence after absolutization
>(and not HTTP equivalence) to determine URI equality for deciding whether
>a multipart/related body part 'matches' an embedded URL.
>
>If the HTTP equivalence rules were supposed to be applied outside of
>HTTP, then RFC 2557 would have to be revised. They're not, and it shouldn't.
>
>
>>  The other point that I made was that "it is ok to do octet-by-octet
>>  comparison if you take into account relative URIs". If an application
>>  wants to do smarter comparison then it is free to do so but you don't have
>  > to.
>
>RFC 2557 doesn't say this, it mandates octet-by-octet comparison.

-- 
_________________________________________
David Durand              dgd@cs.bu.edu  \  david@dynamicDiagrams.com
http://cs-people.bu.edu//dgd/             \  Chief Technical Officer
     Graduate Student no more!              \  Dynamic Diagrams
--------------------------------------------\  http://www.dynamicDiagrams.com/
                                              \__________________________

Received on Thursday, 22 June 2000 15:10:10 UTC