RE: [xml-dev] James Clark: XML versus the Web

["Len Bullard" <cbullard@hiwaay.net>]

Evolution.  Practice and redefinition.  Versioned semantics (one per
enterprise where discipline in a discipline is undisciplined: aka, customer
drift of requirements).

They don't make them unnecessarily complex.  They are.  The problems of
semantic drift among existing forms/messages among business units is
actually profound when your content and data have to cohere among multiple
authorities in a company or other business or social entity.  FB would not
work if the schema weren't locked down in one product.  It is one of the
problems of working with open standards.  No matter how much you try not to,
eventually your implementation and a competitor's implementation have to
drift and datatype drift where that is semantic is huge.

That is/was a strong selling point for markup: a rock reliable
"semantic-free" means to schlep annotated text (scary quotes intentional).  

Adding datatypes to XML has felt to me like a "programmery thing" to do.  In
practice, XML IS a datatype except unlike mapping an int in different
machines, you have to map the other 99 per cent of the application.  

Big task for a syntax-conformant specification.  Too many degrees of
freedom.

len

From: Gavin Thomas Nicol [mailto:gtn@rbii.com] 

I sometimes wonder if people make things needlessly complex in enterprises
app by accident or design... it seems people keep adding and layering
complexity until something breaks... usually in a misguided attempt to
*guarantee* interchanges, reliability, etc. etc.