Having recently seen Danny Ayer's Attention.RDF post, which shows an RDF version of Attention.XML, I thought I would show an Attention.WTF example as a way of showing more about my new WTF (Whoopdedoo Text Format—which I introduced in my previous post).

What follows are the Attention.XML spec examples in OPML and XOXO, Danny's RDF example, and finally my WTF example.

OPML:

<outline type="item"
  permalink="permalink or guid"
  guid="permalink or guid"
  lastread="ISO-8601 date"
  duration="seconds value"
  followedlinks="http://example.com/1 http://example.com/2"
  rel="votelink">

XOXO (XHTML):

<li><a href="permalink or guid" rel="votelink">title</a>
   <dl><dt>lastread</dt><dd>ISO-8601 date</dd>
      <dt>duration</dt><dd>seconds value</dd>
      <dt>followedlinks</dt>
      <dd><ul>
        <li><a href="http://example.com/1">link1</a></li>
        <li><a href="http://example.com/2">link2</a></li>
        </ul>
      </dd>
   </dl>
</li>

RDF/XML:

<VoteLink rdf:about="permalink">
   <title>title</title>
   <lastRead>ISO-8601 date</lastRead>
   <duration>seconds value</duration>
   <followedLinks rdf:parseType="Collection">
      <Link rdf:about="http://example.com/1" at:label="link1" />
      <Link rdf:about="http://example.com/1" at:label="link1" />
   </followedLinks>
</VoteLink>

WTF:

:SiteAttention
title: An Example
permalink: http://example.com/permalink
rel: votelink
lastRead: 2005-04-09
duration: 10 seconds
followedlink: link1
followedlink: link2

:FollowedLinks
id: link1
url: http://example.com/1
label: Link 1

id: link2
url: http://example.com/2
label: Link 2

***

So, the surface-level difference between these examples is the syntax. At this level, WTF might be seen advantageous in terms of there being less clutter: e.g., less mixes of elements and attributes and reduction of the sheer number of field delimiters (i.e., angle brackets, quotes, closing and opening tags).

Going down a level, there are many ways that the WTF version could be normalized differently. For example, a less normalized WTF could contain HTML or RDF/XML in WTF field values or otherwise eliminate the FollowedLinks set, and more normalized WTF could add more sets or otherwise ensure that all possible links (i.e., both permalinks and followedLinks) are represented only once.

It's interesting to consider how each format attempts to mean what it's supposed to mean. While it is never clear enough to see this in RDF/XML, RDF, whatever its syntax, uses one technique—the triple, to make a statement of meaning (i.e., subject-predicate-object).

The XOXO example is interesting in that it capitalizes on the techniques of HTML. So, in this case, the statement of meaning is made to correspond with the meaning suggested by definition terms and lists in HTML. I think this is useful in that it capitalizes on the connections between the data semantics and presentation semantics.

Likewise, the OPML idea of an outline is based on presentation semantics. OPML is a representation created for an outliner tool, and so data in OPML "makes sense" especially when it's thought of in terms of outline tools.

***

WTF is designed so that it can embody many different techniques of representing meaning. This is good and bad.

It's bad in that two documents that look very similar in WTF can mean totally different things (i.e., they can be associated with totally different processing models).

But, I think it's good in that: 1) when a processing model comes first (e.g., OPML or RSS), it's like a shorthand to represent a known structure—which is what I did in my example above. And, 2) it doesn't imply that a common processing model is inherent in the format (which I think is one of the delusions of XML and idealisms of RDF).

Finally, working with WTF, I find myself generally thinking about the implications of the data model rather than the implications of the syntax. And, that is my preference, at least.