Re: [xml-dev] XML Schema as a data modeling tool

I think, two primary models for which XML instances can be used are data (to represent messages, records etc) and documents (for e.g HTML like, where there may be lot of mixed content). If we look at modeling data aspects, I think XML can be used to represent both hierarchic and network structures. XML Infoset inherently looks hierarchic, but it can be mentally and technically be mapped to non hierarchic structures.

Given this nature of XML, I think XML Schemas can be used to model data in general (semantics can be built on top of XML Schema features). But I think, there are other techniques available, which should be used to do enterprise scale data modeling in general (for e.g, ER diagrams, class diagrams etc), which are specialized for data modeling (they provide all the features out-of-the-box and are already very robust).

On Mon, Sep 30, 2013 at 1:57 PM, Hans-Juergen Rennau <hrennau@yahoo.de> wrote:

Hello everybody,

I am interested in approaches to use XML Schema as a technology for constructing data models, which have a scope exceeding the definition of exchange messages (e.g. web service design). One idea may be to use XML Schema for the construction of a "reference model" to which message structures can be related, and I think this is a basic goal of NIEM. Another idea may be to use XML Schema for creating a conceptual data model to which relational database schemas may be related, offering unambiguous points of reference to which semantics and metadata may be connected. So I wonder, is XML Schema also used for enterprise-scale data modelling?

Much promise seems to be implied by the fact that XML Schema describes tree structures, is firmly based on an elaborate information model (Infoset/XDM) and is clearly related to URIs - which makes it a candidate resource for generating various other resources from it with immaculate reliability.

So I would like to learn about approaches of using XML Schema for data modeling purposes - goals, benefits, limitations, case studies, etc. etc. If anybody has any links, information or thoughts, it would be appreciated.

Thank you very much,
Hans-Juergen

--
Regards,
Mukul Gandhi