[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

RE: DNS based URIs that don't imply access method

• To: "'David Hunter'" <dhunter@v...>
• Subject: RE: DNS based URIs that don't imply access method
• From: "bryan" <bry@i...>
• Date: Thu, 25 Jul 2002 17:34:43 +0200
• Cc: <xml-dev@l...>
• Importance: Normal
• In-reply-to: <7AB2229D9557BF4AB41F2BB515D2C3DA37D9CB@torex2>

>At this point, I feel I need to intrude into the thread and ask [again,
I'm
>afraid] for the benefits of using URIs for namespace names?

>The most oft-cited reason for using URIs (especially URLs (and
especially
>HTTP URLs)) is that you can guarantee uniqueness. 
I think when you get into oft-cited reasons you're in an area pertaining
to received ideas and philosophies of usage.

>But is this benefit really worth all of the bother of trying to use the
>little-understood URIs for namespace names?  
Okay here I'm jumping in having gotten very bored with the debate
long-since(really the only debates that can hold my attention long on
xml-dev are the 'debates' where everyone gets together to bash xml
schema) but it seems to me the main problem is the not unreasonable
assumption that they should be dereferenceable. Which leads to the
following:
>The costs (mass confusion, thousands upon thousands of emails to
XML-DEV,
>thousands of person-hours burned on discussing the issue ad infinitum)
just
>seem to vastly outweigh the benefits (easier uniqueness) of using
certain
>types of URI (HTTP URLs being the most common, and most ill-fitted,
>example,
>IMHO).

The question this seems to beg is "why the hell did
they(namespaces-folk) decide to use URI's?!" What it implies to me is
that they wanted to having a structure that was likely to be unique
without saying something like: "you need to generate a GUID" which would
surely have caused worse havoc in the way of thousands upon thousands of
emails to XML-DEV, not require you to make your namespace
dereferenceable but give people a way to build dereferenceability
mechanisms if they felt so inclined which I suppose the whole RDDL thing
was an example of some inclination on the parts of people.

Actually I'm not sure if dereferencing is the proper the term here, RDDL
doesn't 'dereference'
xmlns:nmspce="http://www.namespaces.com/mynamespace" to just one value,
and return "_RNG Schema" or some such value the way that say
dereferencing in Perl might tell you a variable is a scalar, it returns
a number of resources found at that location.

>So what are the other benefits of using URI for namespace names, that
I'm
>missing?

I think the possibility to build a dereferenceable solution is an
enormous value, that is sabotaged by our inability to announce that our
input instances to such a solution, namely namespace-specified instances
of xml, are going to be dereferenceable or not.
Often a sabotaged value is worse than no value at all, as evidenced by
poisoned water sources in a desert.

• References:
  • RE: DNS based URIs that don't imply access method
    • From: David Hunter <dhunter@v...>

• Prev by Date: Re: DNS based URIs that don't imply access method
• Next by Date: XML DOM Level 1 vs. DOM Level 2
• Previous by thread: Re: DNS based URIs that don't imply access method
• Next by thread: reversed referencing and dereferencing
• Index(es):
  • Date
  • Thread