RE: The relentless march of abstraction

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

Umm... if the infoset is hiding information 
not revealed in the syntax, it certainly is for 
the working programmer.   It is a mess to find 
out the data model isn't shared  by  
writing queries or transforms while the 
processor silently appends properties.

The algebra proofs should be hidden.

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

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


-----Original Message-----
From: Tim Bray [mailto:tbray@textuality.com]
Sent: Tuesday, February 27, 2001 11:46 AM
To: Dave Winer; XML-Dev (E-mail)
Subject: Re: The relentless march of abstraction


At 04:38 PM 26/02/01 -0800, Dave Winer wrote:
> I've always felt
>that schema are only needed if you're storing XML content in a relational
>database, but so many applications don't require a relational approach, in
>fact I'd argue that there's nothing about XML that requires a relational
db,
>but of course that's what "most people" use, so put the burden on XML, well
>I don't buy it. If it's not needed and it adds complexity let's us an
>approach that doesn't require it.

Well, lots of other people have uses for schemas outside of the
RDBMS arena.  I agree with Megginson that a lot of people expect
more magic & mojo from schemas than they'll deliver in the real
world.  Still, very useful for industrial language designers; and
I think the datatype stuff will actually turn out to be useful
in lots of places.

Having said all that, I agree that the infoset is a tool for
people building the XML family spec infrastructure, not for
ordinary programmers doing real work.  -Tim