Lists Home |
Date Index |
David Carlisle wrote:
> > It is the _owner_ of the DNS entry that gets to say what the resource
> > is.
> True, but that doesn't say anything about the use of that string as a
> namespace name.
The implication is that the owner of the DNS entry gets to create URIs and
decide how the URIs are to be used. One use of a URI is as a namespace name.
> > Certainly you can abuse the intention of the owner, and if the owner of
> > URI never intended it to be a namespace, then this is simply an abuse.
> "abuse" is an emotive word but apart from your assertion that it is
> abuse, where is it stated that this is abusive. In fact by my reading of
> the namespace rec it is the _intended_ behaviour. That is, one names a
> namespace after some resource that (preferably) you control, by using
> the URI of that resource.
How URIs are intended to be used is said in a number of places, Tim BL's
"design issues", Fielding's REST. It is clear from reading this, that it is
up to the owner of the URI to decide what resource a URI identifies.
> If I give you a (dtd) valid, schema valid, well formed, XML document
> that conforms to the namespace REC that uses a namespace name that
> identifies a resource that isn't a namespace then it is not enough for
> you to tell me it is abusive you have to say what rule has been broken.
I am not sure it has been tested -yet- in court, however I bet that if you
start creating XML Namespaces beginning with the domains
http://www.microsoft.com or http://www.ibm.com and if such namespaces are
used in a way not to the liking of the organization, that you will soon find
a flock of lawyers descending on your abode and causing more havoc with your
life than you care. The ownership of a DNS entry _has_ been tested in court.
Ultimately the rule that is broken is then the rule of law.
In any case I consider it bad manners to use someone else's URIs in ways
that they do not intend, much as you would consider it bad manners if I
started portraying myself as "David Carlisle" to banks, or various
government agencies etc.
> > What results is the same inability to communicate a
> Not at all, the same strings mean different things in different contexts,
> and there are contexts where the string "1 + 1 = 3" would be
> interpreted as true for example.
My point exactly. The problem is that
> <x xmlns="mailto:email@example.com"/>
> is a well formed document conforming to the namespace rec, and the
> namespace of the element in that document _is_ mailto:firstname.lastname@example.org.
> It isn't just me that says it is, XSLT/XPath , a namespace aware DOM,
> etc would all accept this document and confirm the namespace name.
Yup, but if I don't sanction this, I can complain to various authorities
that you are misrepresenting me. People take this seriously.