xml-dev - RE: The failure to communicate XML - and its costs to e-business

["Bullard, Claude L (Len)" <clbullar@ingr.com>]

Precisely.  I came from the markup background 
and then did relational databases.  I needed 
the infoSet to connect the two successfully. 
Suddenly, all the things Steve Newcomb and 
Neill Kipp were telling me made sense.  Without 
a clear definition for abstract types, the 
node abstraction just didn't sink in.  Once 
it did, I realized I could write markup in 
tables without ever having to put a single 
pointy bracket in a field except the one 
that handled the schema mapping on output. 

The handling could be... urmm... abstract. 

For relational programmers, show them a 
table definition of a DTD based on the infoSet types 
and a query that gets a recordset and 
maps it to some XML in which the elements 
and attributes are NOT column and row names. 
Then show them they can use XSLT and skip 
all of that.  Bang!  They get it and can 
start focusing on business layer rules while 
you or someone else starts to designing 
the workflow layer for the sexy looking 
presentation layer your graphics person 
built.

Every now and then this gets pointed out: 
some of us taught ourselves computer science 
after years of training for some other 
profession, in my case, English and music.  
Dvorak is probably in that camp.  Yet we 
are still required to sort out the issues 
and condense them for presentation because 
the Dvorak's set traps on purpose or out 
of misunderstanding.  That is why the 
Dumb articles get listed.  As we make 
our noises here, we condense explanations. 
Those who silently lurk get a good condensation 
eventually, and if they use it responsibly, 
it does a good thing; it dampens uncontrolled 
feedback, and that is how noise becomes signal.

Len 
http://www.mp3.com/LenBullard

Ekam sat.h, Vipraah bahudhaa vadanti.
Daamyata. Datta. Dayadhvam.h


-----Original Message-----
From: Steve Boyce [mailto:SteveB@hbs.com]

> o  XML processor.  Tough.  To really Get It, 
> someone has to become acquainted with the 
> InfoSet.

That's *really* true, that it's hard.  And I think the reason why it was
hard for me, and probably for most people, is coming from a database
background.  In this sense, one starts off mentally mapping XML, schemas,
etc., back into familiar database territory.  The Infoset is alien to this
mental model.   

Thanks for that! My point is that talking about a long list of specs is to
miss the point about selling XML and is in fact falling into the trap set by
Dvorak.  Rather, we need to get across a simple message with concrete
examples about "the world before XML" and "the world after XML."