This topic deals with the Part 8 Template Model, that needs a few small corrections in order to allow instances of TemplateStatement to be the rdf:object for rdf:Property instances, i.e. to be a Role Filler in a template signature.
The story is 7 pages long and can be seen/downloaded from http://www.15926.org/templates/template ... isited.pdf
The MMT members are requested to spend ample time on this, because it is important. See the forum topic of Class Model.
Template Model revisited
Re: Template Model revisited
In summary, the difference between the corrected template model in ISO 15926-8 and the proposed template model, is:
..and a rationale has been given why it was changed like that.
Given that this model only exists (is instanciated) in the Core RDL the change is small.
Also, template instances can now be rdf:type's of (Template, Base Template, Meta Template, Proto Template and Specialized Template). This makes the templates much clearer.
Now the templates do not have to be classified as Core Template, which actually has a vague definition. If the template class exists in the Core RDL, it can be regareded as core (highly generic). The location tells it. But the location has no effect on the data model, which it shouldn't.
If we (want to and) have approved this proposal, I suppose we will have to make a note in the part 8 forum (of this site) to preserve it as a change for its next release.
- The classes starting with "RDL" are out,
- The class Template specialization with its properties hasSuperTemplate and hasSubTemplate is out, and
- Classes Specialized Template and Proto Template with their Statement classes are introduced
..and a rationale has been given why it was changed like that.
Given that this model only exists (is instanciated) in the Core RDL the change is small.
Also, template instances can now be rdf:type's of (Template, Base Template, Meta Template, Proto Template and Specialized Template). This makes the templates much clearer.
Now the templates do not have to be classified as Core Template, which actually has a vague definition. If the template class exists in the Core RDL, it can be regareded as core (highly generic). The location tells it. But the location has no effect on the data model, which it shouldn't.
If we (want to and) have approved this proposal, I suppose we will have to make a note in the part 8 forum (of this site) to preserve it as a change for its next release.
-
- Posts: 283
- Joined: Sun Jan 22, 2012 10:02 pm
Re: Template Model revisited
Since our new member emmagia brought this up in the topic 'acronyms' I want to give this topic of the template model another kick.
We need to discuss the template model as it is defined in Part 8.
To me, as voiced by me already in 2012, that model is flawed.
I propose to discuss http://www.15926.org/topics/template-model/index.htm, simply because we can't delay this much longer.
If anyone is of the opinion that this proposal is incorrect please tell us why.
Under that link you will not find the section of the Part 8 model that is intended to fix the position of the signature roles. It is unclear to me why someone would need this, since:
To answer emmagia's question: this is what you'll find under above link:
We need to discuss the template model as it is defined in Part 8.
To me, as voiced by me already in 2012, that model is flawed.
I propose to discuss http://www.15926.org/topics/template-model/index.htm, simply because we can't delay this much longer.
If anyone is of the opinion that this proposal is incorrect please tell us why.
Under that link you will not find the section of the Part 8 model that is intended to fix the position of the signature roles. It is unclear to me why someone would need this, since:
- * the roles have, in the context of the template, a unique name;
- * when stored in a triple store any sequence is destroyed anyway.
- * the roles in Part 2 relationships and multidimensional object are in alphabetical sequence only by convention
To answer emmagia's question: this is what you'll find under above link: