SourceForge : artf6535: Namespace of Relations

Project: NML-WG Trackers > Schema Progress > View Artifact

Artifact artf6535 : Namespace of Relations

Tracker:	Schema Progress
Title:	Namespace of Relations
Description:	Proposal: Use a full URI in the type of Relations. XML: <nml:Relation type="http://schemas.ogf.org/nml/2012/10/base#isSink"> RDF: @prefix ex: <urn:ogf:network:example.net:2012:> . @prefix nml: <http://schemas.ogf.org/nml/2012/10/base#>; . ex:myPort nml:isSink ex:myLink
Submitted By:	Jeroen van der Ham
Submitted On:	03/14/2012 10:18 AM EDT
Last Modified:	11/30/2012 8:06 AM EST
Closed:	11/30/2012 8:06 AM EST

	Status / Comments		Change Log		Associations		Attachments

Status
Status:*	Completed
Category:*	Identifiers
Priority: *	2
Assigned To: *	Jeroen van der Ham

Comments

Jeroen van der Ham: 11/30/2012 8:06 AM EST
	Action:	Update Closed set to 11/30/2012 Status changed from Last Call to Completed

Jeroen van der Ham: 10/09/2012 12:48 PM EDT
	Comment:	I've just did some global search and replace in the documents and examples in revision 69.
	Action:	Update

Freek Dijkstra: 09/05/2012 9:36 AM EDT

Action:

Update
Description changed from

Proposal:

Use a full URI in the type of Relations.

XML:
  <nml:Relation type="http://schemas.ogf.org/nml/relation/2013/10/isSink">;

RDF:
  @prefix ex: <urn:ogf:network:example.net:2012:> .
  @prefix nmlrel: <http://schemas.ogf.org/nml/relations/2013/10/>; .
  ex:myPort  nmlrel:isSink  ex:myLink

Proposal:

Use a full URI in the type of Relations.

XML:
  <nml:Relation type="http://schemas.ogf.org/nml/2012/10/base#isSink">;

RDF:
  @prefix ex: <urn:ogf:network:example.net:2012:> .
  @prefix nml: <http://schemas.ogf.org/nml/2012/10/base#>; .
  ex:myPort  nml:isSink  ex:myLink

Jason Zurawski: 08/16/2012 7:20 AM EDT
	Comment:	Full URI is fine with me, reading through the changelog on this one the scope has shifted a bit from the original issue :)
	Action:	Update

Freek Dijkstra: 08/07/2012 6:29 PM EDT
	Action:	Update Category changed from Procedural to Identifiers Priority changed from 4 to 2 Status changed from Need proposal to Last Call

Freek Dijkstra: 07/11/2012 8:36 PM EDT
	Comment:	A similar proposal should probably be create for using full URIs in the type attribute for Services (artf6536) and full URIs in the type attribute for Labels.
	Action:	Update

Freek Dijkstra: 07/11/2012 8:32 PM EDT

Action:

Update
Description changed from

Proposal:

Use a full URI in the type of Relations.

XML:
  <nml:Relation type="http://schemas.ogf.org/nml/relation/2016/0/isSink">;

RDF:
  @prefix ex: <urn:ogf:network:example.net:2012:> .
  @prefix nmlrel: <http://schemas.ogf.org/nml/relations/2013/10/>; .
  ex:myPort  nmlrel:isSink  ex:myLink

Proposal:

Use a full URI in the type of Relations.

XML:
  <nml:Relation type="http://schemas.ogf.org/nml/relation/2013/10/isSink">;

RDF:
  @prefix ex: <urn:ogf:network:example.net:2012:> .
  @prefix nmlrel: <http://schemas.ogf.org/nml/relations/2013/10/>; .
  ex:myPort  nmlrel:isSink  ex:myLink

Freek Dijkstra: 07/11/2012 8:32 PM EDT

Action:

Update
Description changed from

We have no preference for XML namespacing of relation names. Roman proposes using namespaces for type attribute <nml:
Relation nml-ext:type="hasExample">


See: https://forge.ogf.org/svn/repos/nml-examples/201203-schema/relation.xml

Proposal:

Use a full URI in the type of Relations.

XML:
  <nml:Relation type="http://schemas.ogf.org/nml/relation/2016/0/isSink">;

RDF:
  @prefix ex: <urn:ogf:network:example.net:2012:> .
  @prefix nmlrel: <http://schemas.ogf.org/nml/relations/2013/10/>; .
  ex:myPort  nmlrel:isSink  ex:myLink

Freek Dijkstra: 07/11/2012 8:29 PM EDT

Comment:

During OGF35 Roman expressed favour for proposal #2 listed below, as that was the way things are done in NM. E.g.

<nmwg:eventType>http://ggf.org/ns/nmwg/characteristic/delay/one−way</nmwg:eventType>;

Jeroen and Freek agreed that this makes things a lot easier.

Action:

Update

Freek Dijkstra: 07/11/2012 8:26 PM EDT

Comment:

At OGF 35 I like a short discussion on identifiers for relations.

Currently, it is just a string in XML. e.g. "isSink", "hasPort"
However, it is a full URL in RDF, e.g.
"http://schemas.ogf.org/nml/2013/10/relation#isSink",
"http://schemas.ogf.org/nml/2013/10/relation#hasPort".

So the translation is to simply prefix the XML-identifier with a fixed (special-purpose) URI. So far I used "http://schemas.ogf.org/nml/2013/10/
relation#".

However, as we found at the previous OGF, this is a problem when a new version of NML is defined in the future, with possibly more relations.
Eg.: "http://schemas.ogf.org/nml/2016/08/relation#isSink". In that case, it is no longer clear which URL should be prepended to the base name ("isSink
").

I can see two solutions:

1. Append the namespace of the element when translating.
E.g.
<nml:Topology xmlns:nml="http://schemas.ogf.org/nml/2013/10/base#">
...
  <nml:Relation type="isSink">
becomes "http://schemas.ogf.org/nml/2013/10/base#isSink"

(thus not "http://schemas.ogf.org/nml/2013/10/relation#isSink")

2. Don't translate relation identifier between XML and RDF, and use full URI in XML (yes, it's ugly, but it's simple):
<nml:Relation type="http://schemas.ogf.org/nml/2016/08/relation#isSink">

Action:

Update

Freek Dijkstra: 04/05/2012 11:51 AM EDT
	Action:	Update Title changed from Relation and Namespaces to Namespace of Relations

Freek Dijkstra: 04/05/2012 11:45 AM EDT

Action:

Update
Description changed from

We have no preference for XML namespacing of relation names. Roman proposes using namespaces for type attribute <nml:
Relation nml-ext:type="hasExample">


See: https://forge.ogf.org/integration/viewcvs/viewcvs.cgi/*checkout*/201203-schema/relation.xml?rev=20&content-type=
text%2Fplain&system=exsy1001&root=nml-examples

We have no preference for XML namespacing of relation names. Roman proposes using namespaces for type attribute <nml:
Relation nml-ext:type="hasExample">


See: https://forge.ogf.org/svn/repos/nml-examples/201203-schema/relation.xml

Jeroen van der Ham: 04/05/2012 11:44 AM EDT
	Action:	Update Assigned To set to Jeroen van der Ham

Freek Dijkstra: 03/28/2012 10:46 AM EDT
	Action:	Update Status changed from New to Need proposal

Jason Zurawski: 03/27/2012 7:30 PM EDT

Comment:

Disregarding my last comment, I was looking at the changelog and not the issue ...

I do not have a strong opinion on typed attributes.  I feel its a bit much given the namespace of the element  and the namespace of the attribute are 
the same 99.9999% of the time.  I suppose in rare instances where we want to use an attribute from a different namespace this will be a positive 
change, but I dont ever see that happening.  Do we have examples of when/why it will happen?

Action:

Update

Jason Zurawski: 03/27/2012 7:22 PM EDT
	Comment:	I don't understand this item, the text looks exactly the same below. Am I missing something?
	Action:	Update

Jeroen van der Ham: 03/14/2012 10:52 AM EDT

Action:

Update
Description changed from

We have no preference for XML namespacing of relation names. Roman proposes using namespaces for type attribute <nml:
Relation nml-ext:type="hasExample">

We have no preference for XML namespacing of relation names. Roman proposes using namespaces for type attribute <nml:
Relation nml-ext:type="hasExample">


See: https://forge.ogf.org/integration/viewcvs/viewcvs.cgi/*checkout*/201203-schema/relation.xml?rev=20&content-type=
text%2Fplain&system=exsy1001&root=nml-examples

Jeroen van der Ham: 03/14/2012 10:18 AM EDT
	Action:	Create

Return