Namespace pedantry

classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|

Namespace pedantry

Daniel Elenius
It seems like Protege makes a difference between the namespace/ontology
URI http://bla.com and http://bla.com#, the second one being the same as
the first, but ending with a '#'. This causes problems. For example, if
I have an ontology which imports http://bla.com#, and a repositry for
http://bla.com, the repository is not used. I think URIs with or without
# should be treated as the same URI. If I remember correctly, that would
be the correct way to do it, but I haven't looked at the URI/XML
namespaces specs recently. In any case, it would be much more
user-friendly. People tend to use or not use the hashes randomly.

Daniel
-------------------------------------------------------------------------
To unsubscribe go to http://protege.stanford.edu/community/subscribe.html

Reply | Threaded
Open this post in threaded view
|

Re: Namespace pedantry

Dan Brickley-2
* Daniel Elenius <[hidden email]> [2006-02-14 14:14-0800]
> It seems like Protege makes a difference between the namespace/ontology
> URI http://bla.com and http://bla.com#, the second one being the same as
> the first, but ending with a '#'. This causes problems. For example, if
> I have an ontology which imports http://bla.com#, and a repositry for
> http://bla.com, the repository is not used. I think URIs with or without
> # should be treated as the same URI. If I remember correctly, that would
> be the correct way to do it, but I haven't looked at the URI/XML
> namespaces specs recently. In any case, it would be much more
> user-friendly. People tend to use or not use the hashes randomly.

RDF's use of XML namespaces has xmlns:foo="http://bla.com/a" and
xmlns:foo="http://bla.com/a#" as different, so I think Protege does the
right thing. I notice that the Metadata tab omits the # in xml:base,
was wondering if that results in the wrong triples ever being generated.
Haven't investigated details tho.

Dan
-------------------------------------------------------------------------
To unsubscribe go to http://protege.stanford.edu/community/subscribe.html

Reply | Threaded
Open this post in threaded view
|

Re: Namespace pedantry

Matthew Horridge

On 14 Feb 2006, at 22:35, Dan Brickley wrote:

> * Daniel Elenius <[hidden email]> [2006-02-14 14:14-0800]
>> It seems like Protege makes a difference between the namespace/
>> ontology
>> URI http://bla.com and http://bla.com#, the second one being the  
>> same as
>> the first, but ending with a '#'. This causes problems. For  
>> example, if
>> I have an ontology which imports http://bla.com#, and a repositry for
>> http://bla.com, the repository is not used. I think URIs with or  
>> without
>> # should be treated as the same URI. If I remember correctly, that  
>> would
>> be the correct way to do it, but I haven't looked at the URI/XML
>> namespaces specs recently. In any case, it would be much more
>> user-friendly. People tend to use or not use the hashes randomly.
>
> RDF's use of XML namespaces has xmlns:foo="http://bla.com/a" and
> xmlns:foo="http://bla.com/a#" as different, so I think Protege does  
> the
> right thing. I notice that the Metadata tab omits the # in xml:base,

Dan - can you clarify where you think the # should be appearing?

> was wondering if that results in the wrong triples ever being  
> generated.
> Haven't investigated details tho.

I don't think that is should - if # is in the base then you should  
get a load of triples that use rdf:about="a", if # isn't in the base  
then triple will be of the form rdf:about="#a" (or rdf:ID="a").

Cheers,

Matthew

-------------------------------------------------------------------------
To unsubscribe go to http://protege.stanford.edu/community/subscribe.html

Reply | Threaded
Open this post in threaded view
|

Re: Namespace pedantry

Stephen Waterbury
In reply to this post by Dan Brickley-2
Sorry to dredge up this thread again, but I'm grappling
with RDF's namespace approach again and need a clarification.
Please see comments below ...

Dan Brickley wrote:

> * Daniel Elenius <[hidden email]> [2006-02-14 14:14-0800]
>
>>It seems like Protege makes a difference between the namespace/ontology
>>URI http://bla.com and http://bla.com#, the second one being the same as
>>the first, but ending with a '#'. This causes problems. For example, if
>>I have an ontology which imports http://bla.com#, and a repositry for
>>http://bla.com, the repository is not used. I think URIs with or without
>># should be treated as the same URI. If I remember correctly, that would
>>be the correct way to do it, but I haven't looked at the URI/XML
>>namespaces specs recently. In any case, it would be much more
>>user-friendly. People tend to use or not use the hashes randomly.
>
> RDF's use of XML namespaces has xmlns:foo="http://bla.com/a" and
> xmlns:foo="http://bla.com/a#" as different, so I think Protege does the
> right thing. ...

I've looked at the spec, but I just want to nail this --
Dan Brickley, I hope you can chime in, because I regard you
as authoritative :) ...

If what you say above is true -- i.e., :

(1) xmlns:foo="http://bla.com/a"
(2) xmlns:foo="http://bla.com/a#"

are different, then let's say each namespace has a "bar" in it ...
obviously, in namespace(2), "bar" is fully qualified as
"http://bla.com/a#bar".  Equally obviously, in namespace(1), we
can't say "http://bla.com/abar", so (since namespace(2) has
pre-empted the "#" delimiter), namespace(1)'s "bar" must become
"http://bla.com/a/bar", right?  (Just to comment on this, I
would have expected the reverse:  "/a" would become "/a#", and
if "/a/" was intended, it would be part of the namespace
URI.)

So the conclusion I would be forced to draw from (1) and (2) is:

xmlns:foo="http://bla.com/a" is the same as
xmlns:foo="http://bla.com/a/"

Is this correct?  (Seems weird to me.)

Thanks for any help!
- Steve
-------------------------------------------------------------------------
To unsubscribe go to http://protege.stanford.edu/community/subscribe.html

Reply | Threaded
Open this post in threaded view
|

Re: Namespace pedantry

Stephen Waterbury
Stephen Waterbury wrote:
> So the conclusion I would be forced to draw from (1) and (2) is:
>
> xmlns:foo="http://bla.com/a" is the same as
> xmlns:foo="http://bla.com/a/"
>
> Is this correct?  (Seems weird to me.)

Oops -- I didn't mean to say that that seemed weird
to me.  What would seem weird would be if

(1) xmlns:foo="http://bla.com/a"
(2) xmlns:foo="http://bla.com/a/"
(3) xmlns:foo="http://bla.com/a#"

were *all* different namespaces, because then I wouldn't know how to
combine (1) with a local name.

There I go answering my own messages again ... :)
- Steve
-------------------------------------------------------------------------
To unsubscribe go to http://protege.stanford.edu/community/subscribe.html