OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help



   Re: [xml-dev] Elliotte Rusty Harold on Web Services

[ Lists Home | Date Index | Thread Index ]

At 8:30 AM -0500 2/3/03, Mike Champion wrote:

>Having done a little of the assigned reading ;-) on philosophy of 
>language over the weekend ... Isn't this the "identity problem", 
>e.g. "am I the same person that I was 30 years ago even though my 
>cells, appearance, ideas, etc. are mostly different"?

I don't think it's the same problem because XML is much more formally 
defined than "me". Furthermore, I am much less fixed in nature than 
XML is.

>I didn't quite follow the reasoning of Wittgenstein here, not 
>surprisingly, but I'm pretty sure that I am the same person that I 
>was, and that XML 2.x or 3.x will be "XML" even if it refactors out 
>some of the stuff that has proven not to work well, or is suitable 
>only to a very limited audience, and perhaps makes it available in 
>an extension language or via a preprocessor rather than in the core 
>definition.  Sure that's confusing to those who want the world to be 
>static, but less confusing than the alternatives IMHO.  Change is 
>going to occur, the question is whether the Powers that Be 
>anticipate and accomodate the forces driving change, or force it to 
>occur via fragmentation.

When and if this happens, it will become important to distinguish 
between XML 1.1, XML 1.0, XML 2.0, etc. just as we have to 
distinguish between K&R C, ANSI C, C-99 and so forth. For now though, 
there's only one XML.

>The world being as it is, I suspect that whatever we "decide" here, 
>the data-oriented applications of XML will drive the evolution of a 
>new species of markup metalanguage that is more optimized to data, 
>e.g. a more programming-like syntax, adding terseness and rapid 
>parsability as design goals, and making it easier to compose multipe 
>documents into a single legal document, etc.  (I suspect that next 
>generation markup language will look more like the RELAX NG compact 
>syntax than XML 1.0, but I digress).

Perhaps. If this happens, I might even write a book or two about this 
NGML (Next Generation Markup Language) in which I would extoll the 
benefits and excoriate the deficiencies of the language, just as I do 
today with XML. I might well compare and contrast it with XML to 
explain to XML developers when they should stick with XML and when 
they should move to NGML, as I do today when writing about Java and 
C. Or I might barely acknowledge the XML heritage if it seems likely 
to more confusing than helpful, and most of my readership doesn't 
really know XML at all, as I do today when writing about SGML and 
XML.  However in either case, I would endeavor not to confuse NGML 
with XML. I would treat it as its own thing with its own strengths 
and weaknesses.

>  Still, I think that alternative syntaxes for the Infoset, a more 
>rigorous definition of what the XML data model really is, and a more 
>layered architecture that allows document and data-oriented users to 
>share what they have in common and fork only on what they don't 
>share, can all live in the "XML" tent quite comfortably, with a bit 
>of mutual respect and flexibility on all sides.

I don't believe that. I specifically don't believe there are 
alternative syntaxes for the Infoset (of if there are, they're not 
XML). I would like a more rigorous definition of the data model, but 
I think the Infoset failure proved it's too late for that. And I am 
opposed to forking anything in the lexical structure of an XML 
document (e.g. document type declarations, CDATA sections, etc.)

>The history of C -> C++ -> Java -> C# is an interesting thing to 
>ponder in this context: to me, it's not self-evident that Java did 
>the right thing in forking rather than calling itself "C++--" or 
>whatever.  It went off by itself, we've had 5 years of language 
>warfare, and now we have 3 contending languages (C++, Java, and C#) 
>that share far more ideas than they disagree over, and whose syntax 
>is easily confused for one another. I would prefer NOT to do this 
>all over again (speaking of real world costs in developer time and 
>productivity!) on the markup language side.

I don't think C++ and Java are all that similar. In fact, I've said 
before, they're about as different as it's possible for two object 
oriented languages derived from C to be. I don't think it's helpful 
to pretend Java is C. I'm glad they have different names. C# is more 
similar to Java, but it's still different enough that it should have 
its own name. I also think its important for there to be a plethora 
of languages. One language does not fit all. Competition is good. I 
do not believe in the centralized "There can be only one" philosophy. 
Perhaps the problem is that the forkers do believe there can be only 
one. Thus rather than inventing their own language that meets their 
needs, they must warp XML to fit. This isn't necessary. New markup 
languages can and will be invented, and they don't all have to be 
called XML.

| Elliotte Rusty Harold | elharo@metalab.unc.edu | Writer/Programmer |
|           Processing XML with Java (Addison-Wesley, 2002)          |
|              http://www.cafeconleche.org/books/xmljava             |
| http://www.amazon.com/exec/obidos/ISBN%3D0201771861/cafeaulaitA  |
|  Read Cafe au Lait for Java News:  http://www.cafeaulait.org/      |
|  Read Cafe con Leche for XML News: http://www.cafeconleche.org/    |


News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS