Hi all, Yesterday we've had a short discussion on the call regarding the topology interchange syntax for NSI. We need to make a decision on this soon, because the implementers need it. Before we get to the options let me make a few things clear: - This is not about the terminology for the topology representation. As a group we've basically reached consensus on the extensions on NML that are required for NSI. I am currently writing these down and will soon send that doc to the list. - This is not about the Topology Service itself. The exact working of the Topology Service still needs to be defined. The requirements of that are currently being written down in another document, after which we can have a coherent discussion about the implementation. - All options listed are valid NML + NSI extensions, and all represent the same abstract models. NML allows for different syntax representations of its model. We have been using different formats for the AutoGOLE demonstration, but this is hurting the implementors since it is not clear which one they should support. With that out of the way, the three possibilities are below: - Plain XML, a direct codification of the NML & NSI concepts into XML. This has not been used so far. - OWL RDF/XML, an RDF representation in XML. This has been used with the DTOX format in the NSIv1 demo, and the NSI format has also been available for the NSIv2 demo, although only used by OpenDRAC I believe. - Notation3, a terser RDF representation. This has been made available since the NSIv2 demo, and used by many different implementations. We should decide on at least one format that must be supported by all implementations. We can also make the other syntaxes optional, if we have a good mechanism to describe the supported options. Jeroen.
Hi On Thu, 24 Jan 2013, Jeroen van der Ham wrote:
Yesterday we've had a short discussion on the call regarding the topology interchange syntax for NSI. We need to make a decision on this soon, because the implementers need it.
I've made an XML version of the nordunet topology for people to see: https://github.com/jeroenh/AutoGOLE-Topologies/blob/nsiv2/goles/nordu.net.xm... I am not entirely sure everything is right, but it should show the structure at least. For comparison: OWL: https://github.com/jeroenh/AutoGOLE-Topologies/blob/nsiv2/goles/nordu.net.ow... N3: https://github.com/jeroenh/AutoGOLE-Topologies/blob/nsiv2/goles/nordu.net.n3 Best regards, Henrik Henrik Thostrup Jensen <htj at nordu.net> Software Developer, NORDUnet
Hi, On 28 Jan 2013, at 11:51, Henrik Thostrup Jensen <htj@nordu.net> wrote:
Hi
On Thu, 24 Jan 2013, Jeroen van der Ham wrote:
Yesterday we've had a short discussion on the call regarding the topology interchange syntax for NSI. We need to make a decision on this soon, because the implementers need it.
I've made an XML version of the nordunet topology for people to see:
https://github.com/jeroenh/AutoGOLE-Topologies/blob/nsiv2/goles/nordu.net.xm...
I am not entirely sure everything is right, but it should show the structure at least.
I just did some syntax fixing on that file, and some further format fixing on the file, I've updated it at the same location. It still does not validate 100% to the NML XML schema, but I have to check with the schema guys whether that is a schema or a syntax problem ;) Jeroen.
Hi Jeroen On Mon, 28 Jan 2013, Jeroen van der Ham wrote:
I just did some syntax fixing on that file, and some further format fixing on the file, I've updated it at the same location.
It still does not validate 100% to the NML XML schema, but I have to check with the schema guys whether that is a schema or a syntax problem ;)
A couple of comments on the last two commits (this is more of an NML thing though and may not be your exact table). The nml:Relation seems like poor XML design instead of using a direct representation. The hasPort should be used to mark that a Node has a BidirectionalPort (I'll try and update the file tomorrow, going home ASAP). The hasPort should probably also be used in .owl and .n3 files, as there is no hasBidirectional relation in the spec (unless I have an old one). Best regards, Henrik Henrik Thostrup Jensen <htj at nordu.net> Software Developer, NORDUnet
participants (2)
-
Henrik Thostrup Jensen
-
Jeroen van der Ham