Re: [xml-dev] Media Types, Purposes, Natures, and XSL Transforms

[Jonathan Borden <jonathan@openhealth.org>]

Eric Hanson wrote:
>
> It seems to me that RDDL is designed to do something, and I'm
> trying to use it to do something else entirely.  Just to recap,
> and correct me if I got anything wrong here, but:
>
> RDDL doesn't mandate the way that resources are described, where
> with Typekit, developing a standard is precisely the point of
> the spec.

We are trying to be flexible because we can't foresee all the ways that 
people might want to describe namespaces. Perhaps RDDL doesn;t capture 
every way of describing a namespace but hopefully it fits the ol' 80/20 
point.
>
> RDDL is human readable, but for Typekit's uses, human
> readability is totally pointless.

The semantic web folks often describe the current web as the human 
readable web and the semantic web as a machine readable web. RDDL is an 
explicit attempt to do both. Having been a sometimes card carrying 
semantic web guy (and man enough to admit this on xml-dev :-) I 
actually do see the value in human readability... this is one of the 
big benefits of XML in the first place.

In any case RDF/XML is also considers human readability as totally 
pointless, and there is a good amount of software that does stuff with 
RDF/XML so I'm not sure that we need more ways of describing resources 
in ways that humans can't read.

You might have more luck talking to the semantic web folks, but I 
should warn you that RDF/XML is often somewhat of a religion over there 
and they will likely see typelit as just YAARXS (Yet Another 
Alternative RDF/XML Syntax). Of course there are a good number of 
semantic web folks who actually hate RDF/XML but that's generally 
because its XML and they would either like to see the web evolve toward 
N3 (TimBLs alternative RDF syntax) or just use s-expressions like any 
real computer scientist ought to be doing. I should warn you that these 
folks have had the opportunity to sharply hone their opinions over many 
decades, and you are unlikely to sway them.

...
>
>> From my perspective, nature and purpose have a general
> usefulness that extends beyond the scope of RDDL.  How about
> abstracting these from RDDL and making them the basis of a
> stand-alone way of describing resources that support XML data?
>
> The first acronym that comes to mind for this is, well, RDF. :-)
> Which makes me think I'm trying to reinvent a wheel here without
> knowing it.  If so, someone place club me with the clue stick,
> but as far as I can tell, this isn't what RDF does.
>
Well RDF isn't human readable and it does describe resources, and 
splinter specs such as RDFS and OWL allow you to specify vocabularies 
in ways that actual software can process. Of course it's probably more 
complicated that what you want, but so is the English language... and 
just like the English language it is being used ... actually if we use 
this analogy, RDF is more like ... say Belgian, but nonetheless it does 
have a population :-)

Jonathan