OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

dita message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]

Subject: Re: [dita] iiRDS and metadata

Re: interoperation


It depends on what you mean by “interoperation”.


For conforming DITA markup it usually means that your DITA specialization is as close as possible to the target standard and that a relatively-simple transform will get you from the DITA markup to the target standard (e.g., using the same local names but having the transform add in missing namespace bindings).


If you go the foreign route then interoperation is pretty much “just chunk your {standard of your choice} here” or take the SVG and MathML route and provide specialized reference elements so you can use foreign markup stored in separate non-DITA documents (and Chris N. has some ideas about making that mechanism more general in DITA 2.x, which I think is a good idea).


Because DITA does not impose any semantics beyond “this is metadata” on the <data> element, there cannot be any semantic mismatch with other standards and, modulo the use of attributes vs subelements, there’s no XML-based structure that can’t be modeled completely using nested <data> elements.


So I don’t see that interoperation in general can be a concern.


Of course, with subjectScheme there’s a little more semantic definition, but even there it’s very generic as regards the basic taxonomic and ontological structures, reflecting universal relationship types and not disallowing any new relationship types.







Eliot Kimber





From: Jim Tivy <jimt@bluestream.com>
Date: Tuesday, February 21, 2017 at 12:46 PM
To: 'Eliot Kimber' <ekimber@contrext.com>, <dita@lists.oasis-open.org>
Subject: RE: [dita] iiRDS and metadata




I like the current section in the spec called DITA Metadata.

We are talking about user defined meta data here – meaning the user, not dita, defines the schema – and DITA defines the meta schema.


Agreed data and data-about are quite powerful.

<data-about href="" scope="external">

<data name="title">The World Doesn't End</data>

<data name="author">

<data name="firstname">Charles</data>

<data name="lastname">Simic</data>


<data name="published" datatype="year">1989</data>

<!-- ... -->



Your discussion is a good overview on the subject of user defined meta data supported in DITA – the spec could develop this idea of user defined meta data with reference to already existing sections on data and subjectScheme.  In DITA there are two ways to represent metadata.  Embedded <data> as you say as well as external – subjectdef and subjectref as per the classification domain.


I think the set of support we have out of the box is good.  Although in theory specialization can define any data model, we have to look at how practical and interoperable this is with other standards like OWL2.

How interoperable is:

1. our out of the box schema and tagging instances with other standards.

2. specializations of DITA meta data elements.





From: Eliot Kimber [mailto:ekimber@contrext.com]
Sent: February-21-17 10:09 AM
To: Jim Tivy <jimt@bluestream.com>; dita@lists.oasis-open.org
Subject: Re: [dita] iiRDS and metadata


DITA provides two completely-general metadata representation facilities, both of which can be adapted to any existing metadata system.


The <data> element allows for arbitrary name/value pairs (simple value), where the name can be anything, e.g.:


<data name=”dc:title” value=”The Title”/>


That alone handles most cases of metadata representation.


<data> can also be nested arbitrarily, so you can represent any structured metadata as well. This should satisfy what I understand to be the iiRDS requirements.


Through specialization you can model pretty much any existing metadata vocabulary with a high degree of fidelity (modulo the use of namespaces, which DITA does not allow in specialized element type names).


The subject scheme facility provides a way to model any possible taxonomy or ontology using DITA-defined syntax, if that’s useful. For example, I generated a subject scheme that exactly reflects the IET’s 25,000-term taxonomy for engineering concepts.


Finally, you can use DITA’s foreign mechanism to directly includes non-DITA, XML-based metadata, e.g., RDFx, etc., if that’s useful.


I don’t think we need to say more than this—anything else would be in the context of specific content and that is outside the scope of the TC’s purview.


For example, if you want to use Dublin Core names on <data> elements you’re free to do so and the TC doesn’t need to have an opinion one way or another. If specific communities of usage want to have opinions that’s entirely different. For example, Publishers might want to establish guidelines for using the various publishing-specific metadata domains with DITA-based content.







Eliot Kimber




From: <dita@lists.oasis-open.org> on behalf of Jim Tivy <jimt@bluestream.com>
Date: Tuesday, February 21, 2017 at 11:54 AM
To: <dita@lists.oasis-open.org>
Subject: [dita] iiRDS and metadata


With regards to metadata and the DITA TC, it would be nice to have some sort of position paper on how DITA supports, interacts with, and relates to meta data in general – with reference to specifics.


The iiRDS spec is one such emerging standard with a proposed iiRDS core.  Dublin Core is another standard which DITA has support for.  As well, there are all the possibilities of subject scheme and embedded metadata.


iiRDS might be better served to adopt the terminology of OWL2 rather than RDFS – as the notations and meta model are more concise.


Currently in manufacturing industries there are other standards that require payment.  Perhaps iiRDs is to establish a free standard to replace RDSPP (below) – the use cases in manufacturing that were referenced are very similar.


Reference Designation System Power Plant (RDS-PP)





[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]