xml-dev - RE: Debug Support in XSL (WAS RE: XML Extensibility / XML Schemas )

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

Thanks Ken.

Yes.  I concur.  The chasm between the typically 
fixed structures common to WYSIWYG, essentially, property 
editors, and the array of semi-programmatic constructs 
required to work with declarative-quasi-procedural 
languages is large.   And knowing your background, 
I am aware just how hard you and your former company 
worked to bridge that chasm.  Even for DTDs which are 
relatively simple compared to the bewildering constructs 
of the XML family of specs, it was a tough nut to crack. 

So yes, without access to the language format, it is hard 
to do serious work.  OTOH, I don't think for experienced 
practicioners this is in question.  Think of how many generations  
of rapid development tools have come and gone, how many SGML 
editors, etc., and in the end, we are still using PFE or see 
the native language editors in the RAD tools.  Caveat vendor.

My intuition is that until we get the fully featured 
design suites promised from the large XML vendors, we
will have a piece meal debug environment.  I have 
recently sat in with several to-the-metal programmers 
who are trying to apply XSL, XSLT, and reasonably simple 
XML:

1.  They know too much to live. (Wouldn't make it past 
the first round of Survivor).  They have been lobotomized 
by the Visual tool sets and the younger ones who don't remember 
C and hand edited make files are struggling.

2.  They laugh when I suggest that using  
Alerts and MessageBoxes might be worthy.  I go back 
three days later and guess what.... Boxes galore.

3.  The concepts and application of namespaces comes hard.  
It defies what they are used to and the implementations 
do things they hate (eg, carrying 
forward the namespace into the output where they didn't 
expect to see it having become convinced they only 
put the xmlns= att for validation).

This is still bleeding edge tech.  The challenge to create 
really comprehensive robust building tools is still unmet 
but I have the same hope as when I saw N&F demoed in Atlanta 
some years ago.  Some of the papers from MS of late on the 
BizTalk orchestration editors are fascinating.  I didn't 
think I would ever see the CreateRFQ, ReplyQ sorts of 
enterprise management tools.  Gad!  The lunatic fringe 
has been validated by a company once considered their 
mortal enemies.  How things change!

DTDs like paper will go away when something as robust and 
easy to use is provided.  Until then, we live in a mixed 
environment.  The convictions of the W3C members are not 
convincing, evidentiary, or compelling.  That Schemas 
are needed is not in doubt.  That we will abandon DTDs 
anytime soon is.  Like HTML, there is too much legacy to lose 
the form therefore compatibility and bloat will be the price of 
evolution, same as it ever was.

Len Bullard
clbullar@ingr.com
http://fly.hiwaay.net/~cbullard/lensongs.ram

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


-----Original Message-----
From: G. Ken Holman [mailto:gkholman@CraneSoftwrights.com]

Personally, I don't, which is why I'm anxious to see what Whitehill offers 
to see the way they support their claims.  It is exciting to think these 
tools are on their way.