Is XML simple? (was Re: Yet Another Xerces/Xalan question)

[Eric Hodges <ha...@swbell.net>]

I think XML is simple.  If you ignore all the other stuff (DTDs, XMLSchema,
XPath, XSL-T, etc.)  XML can be defined with a trivial grammar.  The idea
itself is pretty obvious to anyone that's spent much time converting
different data comm protocols.  Encoding the schema along with the data was
an idea I helped implement at a major check verification company 7 or 8
years ago.

I think Arnaud  Le Hors may have hit the nail on the head with his post on
this topic.

----- Original Message -----
From: "Mike Pogue" <mp...@apache.org>
To: <ge...@xml.apache.org>
Sent: Wednesday, March 01, 2000 2:24 PM
Subject: Re: Yet Another Xerces/Xalan question


> Who said that XML (as defined by the W3C) is simple?  :-)
>
> Mike
>
> > Eric Hodges wrote:

> > The philosophical question:  If XML is so simple, why is DOM so
complicated?

Re: Is XML simple? (was Re: Yet Another Xerces/Xalan question)

[Arnaud Le Hors <le...@us.ibm.com>]

Eric Hodges wrote:
> 
> Is there any historical document of the restraints that were considered in
> the design of the DOM API?  Just out of curiosity, I want to know what they
> were thinking when they did some of it. :)

The only public document in this regard would be the requirements
document [1]. But it doesn't get into any of the gory details that
you're looking for. ;-)

[1] http://www.w3.org/TR/DOM-Requirements/

As far as I know, only the W3C member-only minutes of the face to face
meetings and the mailing list archives contain those. And I guess the
brains of the DOM Working Group members...
One thing I can tell you is that constraints mostly came from people
already having proprietary DOMs and looking for something they could
implement on top of their _existing_ structures.

> It occurs to me that the DOM API
> might be several APIs merged together.

Well, it clearly is. Why do you think there are two basic ways of
navigating the tree: getFirstChild+getNextSibling vs getChildNodes?
Each of them clearly favors one particular implementation. So, given
that nobody wants to give up on his/her preferred navigation model, the
only chance to get to an aggreement is to put both in. Believe me, this
isn't done lightly though. Everybody agrees that a single way of doing
things is best in the first place. But once you've spent many months
arguing on which way to go and everybody's convinced nobody will ever
win the battle, you naturally come to the conclusion that the only
solution, if not best, is to have both.
-- 
Arnaud  Le Hors - IBM Cupertino, XML Technology Group

Re: Is XML simple? (was Re: Yet Another Xerces/Xalan question)

[Eric Hodges <ha...@swbell.net>]

I believe that.  I find the cleanest designs come from two or three people
with similar goals and slightly different perspectives.  Once a committee
gets involved the design usually suffers.

Is there any historical document of the restraints that were considered in
the design of the DOM API?  Just out of curiosity, I want to know what they
were thinking when they did some of it. :)  It occurs to me that the DOM API
might be several APIs merged together.


----- Original Message -----
From: "Arnaud Le Hors" <le...@us.ibm.com>
To: <ge...@xml.apache.org>
Sent: Thursday, March 02, 2000 12:25 PM
Subject: Re: Is XML simple? (was Re: Yet Another Xerces/Xalan question)


> Looking at the DOM in particular, it's actually easy to apprehend the
> difficulty of designing such a standard. We all can design a tree like
> structure, there is really nothing very fancy in this, and most students
> in computer science, if not all, have done it at least once. Now,
> imagine we all get together in a room and decide we must design ONE such
> thing...
>
> I've been involved in the DOM since the beginning and I can tell you
> it's been a very frustrating experience in many cases, however, no
> matter how ugly you may think it is (and I would agree with you on some
> of it!), I claim that nobody could have done better
> _with_the_same_constraints_.
> --
> Arnaud  Le Hors - IBM Cupertino, XML Technology Group

Re: Is XML simple? (was Re: Yet Another Xerces/Xalan question)

[Arnaud Le Hors <le...@us.ibm.com>]

Looking at the DOM in particular, it's actually easy to apprehend the
difficulty of designing such a standard. We all can design a tree like
structure, there is really nothing very fancy in this, and most students
in computer science, if not all, have done it at least once. Now,
imagine we all get together in a room and decide we must design ONE such
thing... 

I've been involved in the DOM since the beginning and I can tell you
it's been a very frustrating experience in many cases, however, no
matter how ugly you may think it is (and I would agree with you on some
of it!), I claim that nobody could have done better
_with_the_same_constraints_.
-- 
Arnaud  Le Hors - IBM Cupertino, XML Technology Group

Re: Is XML simple? (was Re: Yet Another Xerces/Xalan question)

[Pierpaolo Fumagalli <pi...@apache.org>]

Eric Hodges wrote:
> 
> I think XML is simple.

Lexically speaking? It's a markup language... And yes, it's simple...

> If you ignore all the other stuff (DTDs, XMLSchema,
> XPath, XSL-T, etc.)

This is not XML, these are standards based ON xml, but they are not XML
(as the language is defined, of course, place back the DTD in that!).

It's like saying... JAVA is simple (as a language) but the whole
Platform 2 class library is not...

> XML can be defined with a trivial grammar.  The idea
> itself is pretty obvious to anyone that's spent much time converting
> different data comm protocols.  Encoding the schema along with the data was
> an idea I helped implement at a major check verification company 7 or 8
> years ago.

I don't know... The issue was arised by Brian in his keynote to the
XTECH conference... Should be the standards more "implementation
friendly" or more "academically complete" ???

> I think Arnaud  Le Hors may have hit the nail on the head with his post on
> this topic.

I believe Arnaud is right... We need to find a compromise... And, apart
from a couple of doubtful things, the effort seem to have succedeed so
far.

	Pier

-- 
--------------------------------------------------------------------
-          P              I              E              R          -
stable structure erected over water to allow the docking of seacraft
<ma...@betaversion.org>    <http://www.betaversion.org/~pier/>
--------------------------------------------------------------------
- ApacheCON Y2K: Come to the official Apache developers conference -
-------------------- <http://www.apachecon.com> --------------------