The
more basic problem with URL's, URI's, and URN's is
that,
as Michael S-M points out, a domain name isn't permanent,
it
can be abandoned and then picked up by a different party.
While
its highly unlikely that a particular set of schemas, a "standard"
be
spoofed, (much more likely that the validator request
would
return a "404") its entirely possible.
Im
sure this is just a small part of the significant discussions, but its
very irritating.
That
seems to epitomise the W3Cs quagmire created by such irrelevant reductio
ab absurbum arguments for a number of topic, which I believe has been worsening
during the last 12 years. In some circumstances, pragmatism is necessary
(even though XML 1.0 got some things very wrong eg, name collision -
that was rectified: the namespaces concept was introduced).
Isnt
it a problem with the improper association of URLs-URIs-URNs?
The
most important attribute of a namespace identifier is that it is unique.
Authors can guarantee the uniqueness of a URL by registering a domain name
with an Internet naming authority. Then it's the author's responsibility
to make sure that all strings used after the domain name are unique. URNs
work the same way.
To
guarantee the uniqueness of a URN, authors must again register their namespace
identifier with an Internet naming authority. The author is then responsible
for following a scheme for generating unique namespace-specific strings.
Organizations
defining XML namespaces should develop a consistent scheme for creating
new namespace names.
Maybe
Im simplifying or misunderstanding the strict definitions, but heres
a proposition that diverges significantly from the above.
Although
it may be imperfect compared with a strictly-formulated validating or registering
authority, an interim system based on GUIDs (urn:uuid) might have been
widely implemented.
At
any time, the W3C itself (or a commercial organization such as VeriSign)
could have set up an interim system that equated a URN to a unique and
permanent GUID. In the same way that the DNS system operates for URLs,
this (interim) GUID Authority could have maintained a precisely versioned
schema set for each namespace GUID without tying it to a URL locator.
The
404 problem need not ever occur.
The
GUID-URI DNS would always maintain the simple lookup, and it would be
a very modest-sized database. A series of status / validity / last changed
flags could be supplement the system so, redundant GUID-URNs that never
went anywhere could just linger without dying, without detriment.
Since
there is a very large number of unique GUIDs, paradoxically I think that
versioning and authentication is not as difficult a problem as with the
URN, where choosing and usage of URNs is very dependent on the constraints
of human factors.
Ian
Thomas
GeoSciSoft
- Perth, Australia
-----Original
Message-----
From: John.Tucker [John.Tucker@...]
Sent: Tuesday, 22 September
2009 11:00 PM
To: SCISOFT
Subject: Re: [metadata] versions
and profiles of ISO 19139
You
are correct about that ...
One
part of the discussions is around
different
versions or profiles of the same standard ...
neither
the W3C nor anyone else has ever set down a standard mechanism
for
doing this in a namespace ...
in
some situations, an update which "includes" the previous version
and
then merely provides additions and deletions seems most appropriate,
in
some situations, putting different schema versions in the same namespace
seems
most appropriate,
in
some situations, a whole new namespace seem better.
But
there is no "standard" way to do any of these things,
so
schema validating processors simply don't know how to handle
changing
versions.
The
more basic problem with URL's, URI's, and URN's is
that,
as Michael S-M points out, a domain name isn't permanent,
it
can be abandoned and then picked up by a different party.
While
its highly unlikely that a particular set of schemas, a "standard"
be
spoofed, (much more likely that the validator request
would
return a "404") its entirely possible.
-JohnT
SCISOFT
wrote:
>
John
>
I haven't followed these types of discussions, so I'd guess that this
>
suggestion ha been thrashed to death: GUIDs.
>
What's the problem with a unique identifier rather than a registrant
>
authority ?
>
>
IL Thomas
>
GeoSciSoft - Perth, Australia
>
>
-----Original Message-----
>
From: John.Tucker [John.Tucker@...]
>
Sent: Tuesday, 22 September 2009 10:31 PM
>
To: SCISOFT
>
Subject: Re: [metadata] versions and profiles of ISO 19139
>
>
The significance of this issue is
>
that we are all working towards the use of authoritative standards for
our
>
metadata.
>
However, so far, the very nature of the InterNet is that it eschews
>
authority.
>
>
Up through now, there is no way to guarantee the integrity and persistence
>
of
>
any resource.
>
>
Here are the quotes that I selected
>
(with email addresses ... please be nice ...)
>
>
At present, there is (unfortunately!) no service to guarantee the uniqueness
>
and persistence of URN's, Universal resource names. "The fact that
there is
>
no general URN resolution service (besides reading the relevant RFCs) is
>
highly iconvenient and pragmatically it is a killer." - Simon Cox
>
<simon.cox@...>
>
>
"The Namespace Recommendation: It is not a goal that it be directly
usable
>
for retrieval of a schema." -Henry Thompson <ht@...>
>
>
"The only way to ensure safe processing [of different versions](i.e. that
>
respects *all* of the processing strategies allowed for in the XML schema
>
spec) is to be scrupulous about changing the namespace if the schema
>
changes." -Simon Cox
>
>
"But as long as domain names are allowed to be used more than once by
>
different owners, no system of unique identifiers built on the domain name
>
system actually guarantees uniqueness of identifier." - C.M.
>
Sperberg-McQueen <cmsmcq@...>
>
>
-John Tucker
>
> SCISOFT
wrote:
>
>
> John
>
> Unfortunately afaik this list doesn't allow attachments (they get filtered
>
> out?).
>
> I'd like to see the quotes. Perhaps you could paste them in-line into
an
>
> email to the list?
>
>
>
> IL Thomas
>
> GeoSciSoft - Perth, Australia
>
>
>
> -----Original Message-----
>
> From: metadata-bounces+geoscisoft=iinet.net.au@...
>
> [metadata-bounces+geoscisoft=iinet.net.au@...] On
>
> Behalf Of John.Tucker
>
> Sent: Tuesday, 22 September 2009 3:36 AM
>
> To: metadata@...
>
> Subject: [metadata] versions and profiles of ISO 19139
>
>
>
> The <xmlschema-dev@...> mailing list
>
> is a venarable discussion group where some of the
>
> most luninary and faithful contributors to the XML family
>
> of standards still participate and contribute.
>
> During the past 4 weeks or so, people including
>
> Simon Cox, C.M. Sperberg-McQueen, and Henry Thompson
>
> discussed what rules there are and what rules there aren't,
>
> concerning versions and profiles.
>
> Since those of us aiming at ISO 19139 now
anticipate
>
> some new versions and profiles, this subject is extremely
>
> relevant.
>
> Without pretending to have near as much
insight as those
>
> people, I present a few of their quotes for our consideration,
>
> just in case there are people on this list who have not yet
>
> considered the implications. The full discussion should be in their
>
> archives and is well-recommended.
>
>
>
> -John Tucker
Some parts of this message have been removed.
Learn more about Nabble's 