On 4/25/07, <b class="gmail_sendername">Danny Ayers</b> &lt;<a href="mailto:danny.ayers@gmail.com">danny.ayers@gmail.com</a>&gt; wrote:<div><span class="gmail_quote"></span><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Starting from your example:<br><br>&lt;div class=&quot;haudio grouping.reconnection&quot;&gt;<br>&nbsp;&nbsp;&lt;span class=&quot;collaborator hcard fn&quot;&gt;Ammonite&lt;/span&gt;<br>&nbsp;&nbsp;&lt;span class=&quot;work-title&quot;&gt;Reconnection&lt;/span&gt;
<br>&lt;/div&gt;<br>[...]<br>&lt;div class=&quot;haudio grouping.reconnection.sunset&quot;&gt;<br>&nbsp;&nbsp;&lt;span class=&quot;work-title&quot;&gt;Sunset&lt;/span&gt;<br>&lt;/div&gt;<br><br>it could be simplified to:<br><br>&lt;div class=&quot;haudio&quot;&gt;
<br>&nbsp;&nbsp;&lt;span class=&quot;collaborator hcard fn&quot;&gt;Ammonite&lt;/span&gt;<br>&nbsp;&nbsp;&lt;span class=&quot;work-title&quot;&gt;Reconnection&lt;/span&gt;<br>&lt;/div&gt;<br>[...]<br>&lt;div class=&quot;haudio reconnection&quot;&gt;
<br>&nbsp;&nbsp;&lt;span class=&quot;work-title&quot;&gt;Sunset&lt;/span&gt;<br>&lt;/div&gt;<br><br>&quot;grouping&quot; isn&#39;t needed, because class already expresses a kind of set<br>membership. There is still the problem of how to connect the item to
<br>its container, easiest would be:<br><br>&lt;div class=&quot;haudio&quot; id=&quot;reconnection&quot;&gt;<br>&nbsp;&nbsp;&lt;span class=&quot;collaborator hcard fn&quot;&gt;Ammonite&lt;/span&gt;<br>&nbsp;&nbsp;&lt;span class=&quot;work-title&quot;&gt;Reconnection&lt;/span&gt;
<br>&lt;/div&gt;<br>[...]<br>&lt;div class=&quot;haudio reconnection&quot;&gt;<br>&nbsp;&nbsp;&lt;span class=&quot;work-title&quot;&gt;Sunset&lt;/span&gt;<br>&lt;/div&gt;<br><br>However personally I think I&#39;d go a step further and make everything
<br>more explicit, to avoid the parser having to do infer the association:<br><br>&lt;div class=&quot;haudio&quot; id=&quot;reconnection&quot;&gt;<br>&nbsp;&nbsp;&lt;span class=&quot;collaborator hcard fn&quot;&gt;Ammonite&lt;/span&gt;
<br>&nbsp;&nbsp;&lt;span class=&quot;work-title&quot;&gt;Reconnection&lt;/span&gt;<br>&lt;/div&gt;<br>[...]<br>&lt;div class=&quot;haudio reconnection&quot;&gt;<br>&nbsp;&nbsp;&lt;span class=&quot;work-title&quot;&gt;Sunset&lt;/span&gt;<br>&nbsp;&nbsp;Part of &lt;span class=&quot;part-of&quot;&gt;&lt;a name=&quot;#reconnection&quot;&gt;Reconnection&lt;/a&gt;&lt;/span&gt;
<br>&lt;/div&gt;</blockquote><div><br>&#39;Mixing&#39; class and id values seems wrong to me, ie: album&#39;s id=&quot;reconnection&quot; and song&#39;s class=&quot;[...] reconnection&quot;.<br><br>The other problem I can see with this is that in a lot of cases where this non-localized grouping becomes necessary, adding language such as &quot;Part of ...&quot; would be redundant from a human reader point of view; you would essentially be requiring authors to use such lexical conventions in order to make use of grouping when in reality, based on the context of the human-readable content it should be obvious that the song we&#39;re talking about is part of the album we&#39;re also discussing. Requiring the addition of &quot;Part of [container name]&quot; to the human readable content is redundant, however without it, making use of an &lt;a&gt; tag to link the song to it&#39;s album&#39;s markup (which seems to be a very simple and elegant solution to linking a member with it&#39;s container) becomes counter-intuitive: you would need to link the text for the song to the markup for the album, which would be confusing for humans since &#39;songName&#39; would logically be assumed to link to further details/etc for that song, as opposed to the album it belongs to.
<br><br>What about this:<br><br>&lt;div class=&quot;haudio reconnection&quot;&gt;<br></div>&nbsp; &lt;span class=&quot;collaborator hcard fn&quot;&gt;Ammonite&lt;/span&gt;<br>&nbsp; &lt;span class=&quot;work-title&quot;&gt;Reconnection&lt;/span&gt;
<br>&lt;/div&gt;<br>[...]<br>&lt;div class=&quot;haudio reconnection.member&quot;&gt;<br>&nbsp; &lt;span class=&quot;work-title&quot;&gt;Sunset&lt;/span&gt; <br>&lt;/div&gt;<br><br>Very similar to Manu&#39;s previous example of using class values like &#39;
grouping.albumname&#39;, however the suffix approach seems more direct, intuitive and all-purpose...<br><br>I&#39;m not sure if using a &#39;class-suffix&#39; like this is generally acceptable, but it seems to solve most of the problems that have been discussed:
<br><br>1) Uses class instead of id.<br>2) Supports the class attribute being used as an implicit indication of set-membership.<br>3) Indicates the direction of the relationship, ie: which item is the container, as opposed to just using &#39;albumname&#39; as a class value for both album and song, which would ambiguously indicate that both things are part of a group called &#39;albumname&#39;.
<br>4) Is subject-agnostic - this approach can be used for any type of grouping where one item is contained by another.<br>5) Works for N-N relationships, ie: class=&quot;album.member soundtrack1.member soundtrack2.member
&quot;.<br><br>As for the issue of hiding meta-data from the humans, I don&#39;t think that&#39;s what&#39;s happening. If we&#39;re discussing an album, and in the course of that discussion mention a track from that album, I think we can assume that the human consumers of the rendered html will understand the relationship between album and song. Adding meta-data to the markup with a design pattern/uF to indicate the nature of that relationship is simply explaining to the machines what the humans already understand.
<br><br>Am I missing anything?<br></div>