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] Infoset chewing gum (was Re: [xml-dev] linking, 80/20)

[ Lists Home | Date Index | Thread Index ]

Tim Bray wrote:
> This is a popular view, which goes on to complain about the fact that 
> the definition of XML 1.0 was purely syntactical and wish that the data 
> model had come first, with lots of syntax choices.  This point of view 
> is wrong.  XML maximizes interoperability by being defined at the syntax 
> level.  This was a deliberate choice based on a huge amount of experience.
> 
> XLink's definition is very highly interoperable.  Others are arguing 
> that the penalty of having to use namespaces is too high, or that the 
> semantics of actuate are either too little or too much, or that there is 
> no built-in syntax remapping to do retroactive support of existing 
> hypertext dialects.  All these arguments are worth having.  But bear in 
> mind that every time you move your representation away from the 
> syntactic level, you pay a severe interoperability price.  Maybe in the 
> case of XLink the price is worth paying (I don't think so, but can see 
> the other side of the argument). -Tim

starting the discussion about what should have been first, the syntax or 
the data model, probably is only a waste of bandwidth and energy, since 
it is history anyway. and my original argument was not to say that i 
wanted to have an xlink data model because i wanted to have multiple 
syntaxes (this is just an additional benefit, in my view).

i think that once you start doing interesting and diverse things with 
structured data, there must be a common understanding what is relevant 
and what is not, so in summary: what is the essential content of the 
data. and i think this is where the infoset started. xquery would be 
impossible with only the xml syntax. dom had to made the same decisions 
(though it made some of the differently).

my area of interest is linking. and if i want to do interesting things 
with linking information (such as manipulating it through dom/sax, 
styling it through css/xsl, querying it through xquery, accessing 
linkbases through an access protocol), then i better had a data model, 
or i would have to deal with all the syntactic details and peculiarities 
every time i wanted to do these things.

so rather than making an argument that the infoset should have been the 
mother of all xml stuff, i just wanted to point out that a data model 
can be quite useful (as tim himself pointed out in another message), and 
that i would be happy to have one for xlink. if people have bad feelings 
about the infoset (and there seem to be very many of these people :-), 
then what else should the data model look like? i honestly don't know, 
and i think the infoset would be a good point to start.

cheers,

erik wilde  -  tel:+41-1-6325132  -  fax:+41-1-6321035
           mailto:net.dret@dret.net -  http://dret.net/
           computer engineering and networks laboratory
           swiss federal institute of technology  (eth)
           * try not. do, or do not. there is no try. *





 

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

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