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


Help: OASIS Mailing Lists Help | MarkMail Help



   The "what is XML" permathread revisited -- was Re: [xml-dev] XML Binary

[ Lists Home | Date Index | Thread Index ]

On Apr 12, 2004, at 7:58 AM, Elliotte Rusty Harold wrote:

> Different: not OK. If it's not a text format fully compatible with the 
> XML specification, then it's not XML, and please don't use the word 
> "XML" (or preferably the initial letter X) to describe it. XML is 
> text. XML is syntax. XML is not the infoset. The infoset is derived 
> from XML, not the other way around. SAX and DOM are are APIs for 
> processing XML. XML is not a serialization format for SAX and DOM. 
> Alternate representations of the Infoset are not XML. Alternate 
> formats that expose a SAX or DOM API are not XML.

That is a sensible position (and obviously shared by one of the editors 
of the XML 1.0 spec), but it is not Revealed Truth.  I, for example,  
think of "XML" as the whole bloody mess -- syntax, data models, APIs, 
query languages, whatnot.  I agree that adding additional qualifiers is 
the best way to be specific in a context that requires a distinction, 
e.g. "XML 1.x syntax" or "XQuery data model."  I realize that some 
people are quite firmly of the opinion that "XML" means "XML 1.x 
syntax", and that "binary XML" is an oxymoron.  But XQuery (explicitly) 
and DOM, XPath, etc.(implicitly) do not depend on inputs in XML syntax.

A religious debate over the true meaning of "XML" seems doomed to cause 
far more schims within the XML community and confusion in the world 
outside XML geekdom than any increase in semantic precision would gain 
anyone.  I'm comfortable saying that XML is sortof a trinity -- "one 
technology in three persons" : syntax, data model, manipulation 
language -- but recognize that these are essentially vague and 
interdependent concepts that will probably never be clarified to 
anyone's satisfaction. Likewise, there are alternatives on any one axis 
so long as one doesn't get too far away from the middle on all axes as 
once.  A bit uncertain, perhaps, but such is life.

In other words, IMHO "XML" is a fuzzy set, not an Aristotelian 
category.  Trying to define it or rigidly constrain it --"XML is not a 
serialization format for SAX and DOM. Alternate representations of the 
Infoset are not XML. Alternate formats that expose a SAX or DOM API are 
not XML. " -- is an interesting academic (or mailing list) exercise, 
but let's not get dogmatic about it ... and leave the flamethrowers at 

[Aside, speakiing of a "trinity": The challenge of defining the Trinity 
apparently keeps certain people occupied fighting the numerous heresies 
still today.  http://apologetics.johndepoe.com/trinityfalse.html  is 
the top Google hit for "trinity heresies" and contains the fascinating 
tidbit that those who believe that the Trinity is one Being in three 
persons are orthodox, but those who believe that it is one person with 
three ways of working are guilty of the Modalist heresy.   I remember a 
college history teacher who had a bit of fun tripping up the professed 
Christians in the class with their inability to distinguish the 
Revealed Truth from the various historical heresies.  "Sorry, you would 
have been burned at the stake for holding that opinion."   I wonder if 
the believers in the Revealed Truth in the class remember that session 
as clearly as I do 30 years later... I certainly came away with a 
lifelong non-interest in making rigid distinctions about intrinsically 
fuzzy concepts]


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

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