[Sigia-l] That stuff we do, ontologically speaking

[Alexander Johannesen]

On Feb 11, 2008 12:57 AM, Eric Scheid <eric.scheid at ironclad.net.au> wrote:
> http://www.wowgemfinder.com/cut_gems/list/matches-meta-socket
> http://www.wowgemfinder.com/cut_gems/list/matches-meta-socket+healing
> http://www.wowgemfinder.com/cut_gems/list/healing+intellect
> http://www.wowgemfinder.com/cut_gems/list/intellect+horde
>
> and so on. Lots of different URLs, all the same resource, just different
> filtered views of that resource.

They are not actually all the same resource as each URL is a resource
in itself. Depending on how you design the backend they can be proxies
for another specific resource (like the generic search URL
http://www.wowgemfinder.com/cut_gems/list), and you can represent this
within your HTTP headers. When we talk about the various
representations of a resource, we need to look in the HTTP headers to
see what is represented, and we need to send our request with what we
want.

For example, something like
http://www.wowgemfinder.com/cut_gems/list/intellect+horde when viewed
through a browser (which accepts the default representation, on the
web most likely some HTML page) gives us "Content-Type: text/html", so
if your browser knows what to do with that content-type (which they
all do), then you get a web page rendered. But any resource can
represent any number of content-types, and if the server says it can
represent this resources as a PDF, for example, you can request the
PDF version of that list instead of the HTML version. That's the real
beauty of what's called "content negotiating" as applications can
choose the content-type representation that they themselves support.
If you create versatile systems, you can have your resources be
represented in APP, SOAP-proxy, plain XML, HTML, text, RSS, Atom, GIF
pictures, and so on, and your users (applications) use what they
themselves find to match the application requirements. This has huge
benefits for resource-orientated solutions, and, in many ways, subject
centric thinking.

A person might use his browser to view a blog in HTML by surfing to
it, but he can point his RSS reader to the *exact* same URL to read it
in Atom format, or point his TM application to the *exact* same URL to
get a Topic Maps representation of the same. Content negotiating isn't
all that much used, but I suspect with the growing popularity of
RESTful systems this will soon change.


Alex (now much colder than 3 weeks ago)
-- 
---------------------------------------------------------------------------
 Project Wrangler, SOA, Information Alchemist, UX, RESTafarian, Topic Maps
------------------------------------------ http://shelter.nu/blog/ --------