An article (Dubliners) [1] about the Dublin Core (Dublin Core Metadata Initiative) [2] at CONTENU.nu (contenu.nu: a content consultancy) [3] got me thinking about the problem of indexing weblogs [4]. The major problem is that there is no semantic markup to include meta-information in the body of a webpage. Sure, you can include meta-information in the <HEAD> section, using both <META> and <LINK> tags, and that's fine when the page in question is about a single topic.
But a weblog has several, mostly unrelated entries on a single page, with the rare weblog having several nearly article-length entries on the main page (and by extension, the archive pages). Google [5] indexes these pages as if it were on a single topic and as a result, you get fodder the Disturbing Search Requests [6].
There are heuristics that can be used to index a weblog page, but it would be nice to have some defined way to mark individual entries, with the ability to include meta-information for each entry. I had intended for my software [7] here to build up the <META> tags (since I do include keywords/classification for each entry I write) and while that may be viable for up to a weeks worth of entries on a page, it starts getting silly for a month, and for a whole year? It's just not practical.
But from the Dublin Core article, I ended up at the W3C (World Wide Web Consortium) [8] site and came across XHTML 1.1 (XHTML 1.1 - Module-Based XHTML) [9] (eXtensible HyperText Markup Language), which is still being worked on, but (and this is the exciting part here) this version of XHTML (eXtensible HyperText Markup Language) can be extended! (unlike XHTML 1.0, even though the name says it's extensible) It's completely modular so new variants of XHTML (for example, it can be extended to MathML (W3C Math Home) [10]) can be constructed from bits and pieces of existing XHTML modules.
So in the future, it may be possible to extend XHTML to include meta-information in the middle of a page, instead of just in the <HEAD> section (sorry, <head> section—XHTML uses lower case for tags). So instead of having to parse code like:
<h3><a class="local" id="2002/07/16.1" href="/2002/07/16.1">The Ins and Outs of Calculating Browser Usage</a></h3> <!-- programming, statistics, web browsers, web log files --> <p> I spent the past few ... <h2><a class="local" id="2002/07/14" href="/2002/07/14">Sunday, July 14, 2002</a></h2> <h3><a class="local" id="2002/07/14.1" href="/2002/07/14.1">Probability</a></h3> <p> ...
It can, instead, have an eaiser time with:
<entry>
<head>
<meta name="keywords" content="programming, statistics,
web browsers, web log files">
<link rel="permalink" href="/2002/07/16.1">
<link rel="next" href="/2002/07/17.1">
<link rel="previous" href="/2002/07/14.1">
</head>
<body>
<p>
I spent the past few ...
</body>
</entry>
<entry>
<head>
<meta name="keywords" content="daily life, web pages, home pages,
six degress of separation, Tom Hoylrod">
<link rel="permalink" href="/2002/07/14.1">
<link rel="next" href="/2002/07/16.1">
<link rel="previous" href="/2002/07/13.1">
</head>
<body>
...
</body>
</entry>
[1] http://www.contenu.nu/article.htm?id=1224
[6] http://searchrequests.weblogs.com/
[7] https://boston.conman.org/about/