RE: Schemas Article

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

Of course not all.  But would this communication 
go anywhere had the mail standards not been created?

It comes down to who gets to choose the choices and 
when.   Communication is about choice but choice and 
consequence are coupled.  

XML Schemas are baroque, no doubt about it.  On the 
other hand, given that we have very little real 
wide experience with them yet, we have to give that 
time to see what the results will be.  I remember 
a very famous designer telling people he could not 
conceive of people bothering to type in tags but 
they do it.  Humans handle complexity one simple 
piece at a time and then it doesn't look complex. 

If all TREX does is validate, how will that be 
any better than a DTD?   If RDF enables semantic 
interpretation but is a bear to learn and work 
with, who will bother to encode metadata?  Everyone 
stops thinking something is "too hard" once they 
learn it.  So maybe the pain has slipped back 
to the programming task.  XML was designed to 
make life easy for the programmer at the cost 
of the author.  Now the pendulum swings. 

XML Schemas will take a year to two years to 
be fully deployed and only after that will we 
know how much pain there is.  What I do know 
is that for all the simplicity XML was supposed 
to engender, what we have is now a heckuva lot 
harder than SGML.  Why?  Lots more features.
Why?  James demonstrates specious logic.  The 
ISO standards groups were very small groups. 
Attendance was usually no more that eight to 
a dozen people and usually the same people. 
It is the W3C working groups that are very 
large groups.  Large groups do create lots of 
features but not because people want satisfaction 
but because they all want to do different tasks.  
If you want small design groups, carefully 
controlled features, and the *right people*, 
found a company and write specification for 
applications.  Proprietary.  Then when it 
succeeds, submit is for a PAS.  Works for Adobe.

But get the requirements right.

If you want OOP, do OOP but that is not markup. 
If you want markup do markup, but that is not OOP.
If you want both from the same language, you 
get a very baroque result.

The features of SGML were carefully 
controlled but the standard was old by the 
time the web got to it, so yes, it was feature 
rich at that point.   SGML DTDs did not fail.  They 
do what they are designed to do reasonably 
well for the user they are designed for.  They 
aren't designed to make a programmer's task 
easy.  It is the web work, with all its whining 
about simplicity that has become baroque because 
its features are not carefully controlled.  It 
won't stay simple.  It never does.

But now we have it and them, and if what comes 
next is painful, well, blame the W3C and the 
XML designers.  Them is us.   

>Communication without contracts is what has enabled the progresses made
>by the humanity (sorry for the lawyers who read this sentence)...

Dead wrong.  Communication with contracts 
is how progress is made.  It keeps the mammals from 
devolving from parlor gossip to swords.

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

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


-----Original Message-----
From: Eric van der Vlist [mailto:vdv@dyomedea.com]
Sent: Wednesday, June 06, 2001 5:00 PM
To: xml-dev@lists.xml.org
Subject: Re: Schemas Article


"Bullard, Claude L (Len)" wrote:
> 
> Or maybe the SGMLers were right all along
> except for requiring them in every case.
> Well-formedness is a coding freedom not
> necessarily extended to the user of the
> coded message but even the SGMLers know
> it isn't required to read the document
> in every case.
> 
> Somewhere the contract for communicating for
> the duration should be expressed.  

Not for every kind of applications.

Would you have written this mail if you had needed to write a schema to
formalize a contract making sure the every reader in the world will
understand each of the words in the same way you've wanted to express
them?

Communication without contracts is what has enabled the progresses made
by the humanity (sorry for the lawyers who read this sentence)...

> Lighter or
> heavier makes no difference to the necessity
> to trust AND verify except where trust is
> proven to work and you can afford the occasional
> defection.
> 
> Besides, XML Schema is just an
> application of XML.  People are free to
> ignore it.  They just aren't free to improve it.

Today, yes it's still the case.

Will it still be if W3C XML Schema becomes a foundation of XML as shown
by Tim Berners-Lee?