Re: [xml-dev] Need a language whiz: An XML Schema "specifies" howdata is

["Norman Gray" <norman@astro.gla.ac.uk>]

Greetings.

On 4 Jan 2018, at 14:46, John Cowan wrote:

> But it
> seems to me that this is only half of assessment: if a document is 
> valid,
> you may still decide to reject it for any of a variety of reasons:
This, many times!

I think people often become obsessed by XML validation, apparently 
forgetting that in any non-trivial document/workflow there will be 
multiple layers of validation of the type which John illustrates, which 
cannot reasonably be articulated with any schema, and in comparison to 
which schema validation is merely an initial sanity check, or possibly 
contract check.

If, for example, a postal code has a mistaken format, and this is caught 
by a schema check, then the resulting error ('computer says no') is 
likely to be much less intelligible to the 'sender' than if the error is 
caught at a higher level, where there is more semantic context.  This is 
a type of 'brittleness', which in my experience is often associated with 
complicated schemas.

It may therefore be a wise general position to avoid designing intricate 
schemas as much as possible, and to perform accept/reject decisions 
relatively late (I can easily think of cases where this would not be 
appropriate, of course).

Best wishes,

Norman


--
Norman Gray  :  https://nxg.me.uk
SUPA School of Physics and Astronomy, University of Glasgow, UK