Comments on: alien vs predator: www-style

By: ed

ed — Wed, 23 Dec 2009 15:44:25 +0000

@danbri thanks for the pointer to DataRSS from Yahoo. I had seen that before but failed to make the connection between it and some things I had been thinking about at the time. Yes, I guess I’m an RDFa optimist — but ultimately I’m an optimist about the RDF data model, and the Web :-)

By: danbri.org/

danbri.org/ — Wed, 23 Dec 2009 08:33:38 +0000

See also DataRSS from Yahoo SearchMonkey folk –

http://developer.yahoo.com/searchmonkey/smguide/datarss.html
http://developer.yahoo.com/searchmonkey/smguide/understand_datarss.html

This uses rdfa. Are you an rdfa optimist?

By: me.yahoo.com/rossfsinger

me.yahoo.com/rossfsinger — Fri, 06 Nov 2009 16:01:47 +0000

I think you’re generally right about Atom being about as good of a compromise as you can get; it’s the same basic reason I’ve been pushing for it for library data.

The danger is more what appears between the tags. Giving a 3GB blob of lots of discrete data in some arcane format a URI and saying, “here, we’ve made it available! Just check out the Atom feed!” isn’t really a huge improvement over just throwing them into a web-accessible directory. It is also the sort of thing that really undermines the perception of Atom as a powerful and flexible format for making things available on the web.

By: sgillies.net/

sgillies.net/ — Wed, 04 Nov 2009 09:09:58 +0000

Yes. This feels like one of those situations where worse (Atom) is better.

By: ed

ed — Wed, 04 Nov 2009 09:02:24 +0000

@johncowan "I will call him ... mini-RDF" :-) Seriously though Atom already seems to be a suitably good graph serialization for a graph of web resources doesn't it? It lets someone say here's a set of resources with URIs (atom:entry, atom:id), and here are the relationships between those resources and other resources (atom:link), and here's some core metadata about the resources (atom:title, atom:summary, etc), and here's some space to say whatever else you want about the collection of resources or the resources themselves (use of extensionElement in atom:feed and atom:entry) There are a few proposals on the table for bringing the full expressive power of the RDF model to Atom: AtomTriples from Mark Nottingham and Dave Beckett, and the work the Open Archives Initiative Object Reuse and Exchange group did on their Atom serialization for graphs of web resources. I don't see people clamoring over themselves to use them however. But (if you are being serious, heheheh I have my doubts) perhaps you see something simpler that can be done?

By: johnwcowan

johnwcowan — Tue, 03 Nov 2009 22:51:09 +0000

So invent an RDF serialization that is Atom-compliant. It’ll be a huge win.