cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Stefano Mazzocchi <>
Subject Re: cvs commit: xml-cocoon/xdocs/drafts sitemap-working-draft.xml
Date Fri, 16 Jun 2000 22:43:25 GMT
Ross Burton wrote:
> > > Maybe it's because I only got out of bed 20 minutes ago and am still
> > > asleep - but CC/PP?  I'm confused... more coffee.
> > CC/PP -> Composite Capabilities/Preference Profiles
> > This is a very recent W3C note.
> > C'mon, people, it's the bleeding edge, you can't afford to miss anything
> > that happens on W3 :)
> > No, seriously, I'd love to see something like CC/PP happen, even if I'm
> > _seriously_ concerned about its complexity given the proposed
> > architecture requirements.
> It's a bit, well, huge and complex from what I have read of the spec.  The
> idea is great but I think RDF is a slight overkill - the Golden Hammer
> antipattern I feel...

Well, the situation they are trying to solve is incredibly complex: how
do you express the capabilities of your user agent without creating a Mb
long HTTP request? how do you make this flexible enough to stand the
variety of user agents and devices that will emerge?

About the golden hammer.... well, RDF is a graph formalization plus a
way to serialize it using XML well-formed syntax.

When using RDF, you are simply saying 

 somebody ---does----> something

not much more. Then you have RDFSchema that adds

      actor                 --- plays --->                  play
(which is a person) (which is a way of performing) (which is a

this is nothing strictly related to XML, its about "describing data with
data", in short "metadata", in practice an architecture to define
"semantic networks". The XML relation is the particular XML encoding
syntax they propose, but it's clearly written that RDF could be
serialized in whatever way you find it useful (of course, since you have
XML it makes it a perfect choice for encoding syntax, even if it gets a
little verbose)

What you are doing with CC/PP is "describing the capabilities of your
user agent" so that servers (such as Cocoon) might understand what to
serve them. It's already difficult today with browsers with different
capabilities, tons of plugins, handheld devices and so on. 

Consider a world where you fridge connects to the grocery store and
orders food for you.

Now, suppose that CC/PP defines a semantic network where, for example,
there are things like "SVG", "PNG", "XSL:FO".... then your IE8.0 sends
its CC/PP-encoded request indicating that it supports "Macrosplit
Formatting Object Language 4.0".

Today, if you don't know about a language, you simply can't do anything.
But if you have a semantic networks that tells you that 

             IE8.0          --- supports --->     "Macrosplit FOL 4.0"
 (which is a desktop browser)                (which is an XSL:FO

then Cocoon is able to come close enough sending FO content. Sure, it
won't be using the fancy <banner> or <blink> features that this new
language implements, but it's better than sending "sorry, don't know
what this FOL 4.0 is".

The Golden Hammer is used when you have a technology that is not well
suited for your stuff and you use it anyway because of symmetry or other
purely esthetic reasons.

Here, I see a perfect need for reusing the RDF ideas, since just like
you do with XML for parsers, an RDF-based semantic processor could be
used to extract information from documents as well as understanding what
your user agent supports.
Stefano Mazzocchi      One must still have chaos in oneself to be
                          able to give birth to a dancing star.
<>                             Friedrich Nietzsche
 Missed us in Orlando? Make it up with ApacheCON Europe in London!
------------------------- http://ApacheCon.Com ---------------------

View raw message