Re: [syndication] Re: Thoughts, questions, and issues.

[Ian Graham <ian.graham@utoronto.ca>]

On Wed, 16 Aug 2000, Aaron Swartz wrote:

> Ian Graham <ian.graham@utoronto.ca> wrote:
> 
> > To my mind the goal
> > should be a design such that, if you don't know the namespace-extended
> > portions, you can still usably process the 'core' RSS (or whatever) data
> > in the message - and, moreover, that the core should be really obvious
> > and simple to use/understand.
> 
> I think this was always the idea -- the RSS core allows description of a
> "channel" with multiple "items" each of which have a "link". That's all it
> says, but all RSS files should have that. You can add whatever you want on
> top of that, but at least that will tie them all together.
> 
> > In this sense, I see work on understanding the correct 'modules' to build
> > [as in http://purl.org/rss/1.0/modules/] as something that should happen
> > before RSS is redrafted....
> 
> I don't quite understand. As I see it, modules are something that allow
> people to extend onto RSS, not something that fits into the core. I think
> your proposals would be great for an RSS 2, something where we can start
> fresh and throw away the Netscape baggage.

It may be that I am wanting RSS 2 already ;-).  In my case I mean 'module'
in the sense of a discrete, reusable block of markup that is, at its base,
defined in the RSS syntax.  I would argue that the core 'parts' of
syndication messages can be cleary separated and then expressed as
distinct markup parts. Roughtly this could be expressed as (dropping all
resemblance to RSS for now):

<message>
   <accessInfo> <!-- how to access this message --> </accessInfo>
   <syndicationrules> <!-- start/stop date. etc.--> </syndicationrules>
   <descriptor> <!--keywords, categories, etc. -->  </descriptor>
   <metadata> <!--whencreated, by who, contact info --> </metadata>
   <messagedata> .... actual content in whatever format  </messagedata>
   <message> ... more messages inside ... </message>
   <message> ... more messages inside ... </message>
</message>

Where 'message' is simply a package of syndicated data of some sort.
Within each component (accessinfo, etc.), namespace(s) could be used to
extend from a core syntax. 

My sentiments are that namespace-based modules are best when you are
intertwining different existing data models or schemas. Right now (as I
see it) we're trying to _understand_ the syndication data model, so it
seems we should be trying harder to get to this understanding and 
build it into the markup model, and not assume that namespaces and
RDF will solve the problem for us. They may (and probably can) do so,
but the result will likely be far messier and more cumbersome than
we want.

How would this related to RSS?  Well, that's what I'm trying to think
through.....

> Does this make sense?

Yes, of course ;-)  I think it's that I want to do something more than
the RSS 1.0 draft was intended to accomplish.

Ian
--
Ian Graham ..........................  http://www.utoronto.ca/ian/
i a n   d o t   g r a h a m    a t    u t o r o n t o   d o t  c a 
Tel: (416) 978-4548 .......................... Fax: (416) 978-7705