Re: [xml-dev] CCTS NDR tools for generation of schemas

[David Carver <d_a_carver@yahoo.com>]

bryan rasmussen wrote:
> hi,
> someone I'm working with will soon be arguing for wholesale adoption
> of the CCTS NDR and using as their argument that users will be able to
> generate schemas just by writing models as opposed to having to deal
> with the complexity of schemas.
>
> This seems pretty wrong headed to me, but I must admit I am not
> knowledgeable enough about CCTS NDR and the modelling thereof to know
> how advanced any applications for doing this modelling are. I am going
> to assume not really too advanced.
>
Bryan, there is at least one tool that will generate most of the CCTS 
NDR, but it generates a narrowed profile of the NDR for XML, not the 
full selection.   You can take a look at GEFEG FX, and make your own 
opinions on how well it does.

Unfortunately, as more tools come out I think this will be the case, in 
that they will generate profiles of the NDR instead of supporting all 
the rules.

> An especial quibble I have is, yes it may generate schemas from models
> but will schemas generated from models be able to work in a broad
> cross section of processors. It seems somewhat wrong headed to expect
> automatic execution of schemas from models when processors for schemas
> often have a lot of incompatibilities.
>
The main compatibility problem that you may experience is in how imports 
and includes and namespaces are supported.  Also, any tool that 
generates schema should at the minimum test the schemas it genrates 
against the most widely used schema parsers.  Unfortunately, most tools 
just generate the schema and say that it works.

> I am also more apt to think that writing code is preferable to
> generating code, partially caused by the experience I have that making
> the models is as complicated as writing the code in a lot of
> instances.
>
I tend to agree, as I've taken an agile aproach to working with 
schemas.   I tend to use UML and other modeling tools as more of a proof 
of concept design, and then write unit tests and schemas to implement 
that design.   In a data driven modeling world I think we are still a 
long way off from truely realizing Model Driven Development, unless you 
have the ability in a tool to tweak the way the Platform Specific Model 
is generated, you will always be at the mercy of how the tool authors 
profiled their platform specific implementation.

Even what is modeled, may not be the best way to implement in the real 
world.   What looks good in a diagram may have performance issues when 
implemented.

Dave




> Cheers,
> Bryan Rasmussen
>
> _______________________________________________________________________
>
> XML-DEV is a publicly archived, unmoderated list hosted by OASIS
> to support XML implementation and development. To minimize
> spam in the archives, you must subscribe before posting.
>
> [Un]Subscribe/change address: http://www.oasis-open.org/mlmanage/
> Or unsubscribe: xml-dev-unsubscribe@lists.xml.org
> subscribe: xml-dev-subscribe@lists.xml.org
> List archive: http://lists.xml.org/archives/xml-dev/
> List Guidelines: http://www.oasis-open.org/maillists/guidelines.php
>
>