xml-dev - Re: XML Schemas: Best Practices

[Jon Cleaver <j.cleaver@eris.dera.gov.uk>]

Roger, All,

There appears to be a convergence of thought that the design of an XML
schema is somewhat analogous to the creation of an Object Oriented Design
for a piece of code. There are instances where you want to maximise
component reuse and instances where you want data to be private. There are
also instances where you want to combine both of these.

What are the guidelines for creating private and public members of an OO
class? Is there a best practice for this - I assume there must be one as
part of an OO design course. Personally, I just decide which components are
needed to be viewed and which I would rather have left alone.

It seems to me that a fair chunk of this thinking will have already been
tackled by our OO design brethrin and there will be a published set of
guidelines. It is just a matter of finding them and their application to
this problem.

Hope this adds to the discussion

Cheers

Jonathan

"Roger L. Costello" wrote:

> Caroline Clewlow wrote:
> >
> > The third approach seems to bear a resemblance to the reasoning behind
> > object oriented approaches to programming.  i.e. re-use,
> > encapsulation, and data hiding ( namespace hiding in this case ;).
>
> Excellent observation.  I will note it in the online version of the Best
> Practices document.
>
> > To reflect this it may be an idea to name it the Object Design
> > approach
>
> I will mark this down as a possible name for the Third Design.
>
> > The choice between the three approaches seems to again come down to
> > a style driven decision.
>
> Oh my, I do hope that it's not a matter of "style".  "Style" seems to
> suggest "by intuition", or "by personal preference".  My hope is that we
> are developing guidelines to empower a schema designer can make
> intelligent decisions based upon a concrete decision space (i.e., all
> the design alternatives are spelled out, and the pros and cons of each
> design alternative is clearly understood).
>
> However, your point is well taken.  We have not clearly spelled out the
> implications of the First, Second, and Third design.  Please keep us
> focused until you (everyone) feel(s) comfortable that it is no longer a
> "style" decision, but rather an "engineering" decision.  Thanks! /Roger

--
Programming today is a race between software engineers striving
to build bigger and better idiot-proof programs, and the Universe
trying to produce bigger and better idiots.  So far, the Universe is
winning.