Curiouser and Curiouser!

Curiouser and Curiouser! on semantic-web http://matt.blogs.it/ RSS feed for topic semantic-web Copyright 2006 Matt Mower Squib/0.1 self@mattmower.com self@mattmower.com en-gb Semantic searches for bloggers Mon, 24 Jun 2002 11:55:21 +0100 <A href="http://www.idlewords.com/semantic/">Semant-O-Matic</A> may be just the thing. Its author, Maciej Ceglowski, calls it a Blog semantic search engine. [<A href="http://www.scripting.com/">Scripting News</A>] <BLOCKQUOTE dir=ltr style="MARGIN-RIGHT: 0px"> » Cool</BLOCKQUOTE> http://matt.blogs.it/entries/00000035.html Classification schemes Thu, 27 Jun 2002 14:56:21 +0100 <A href="http://www.steptwo.com.au/columntwo/archives/000030.html">Classification schemes</A>. I've just stumbled across a paper on faceted classification of information, which talks about applying multiple sets of indexes to [<A href="http://www.steptwo.com.au/columntwo/">Column Two</A>] <BLOCKQUOTE dir=ltr style="MARGIN-RIGHT: 0px"> » Note to self: Read this paper.  </BLOCKQUOTE> http://matt.blogs.it/entries/00000054.html Web Ontology Thu, 08 Aug 2002 22:55:13 +0100 <A href="http://www.iaslash.org/node.php?id=4576">OWL Web Ontology Language Working Drafts Published</A>. From W3C news releases: The Web Ontology Working Group has released three first Working Drafts. The <A href="http://w3.org/TR/2002/WD-owl-features-20020729/">Feature Synopsis</A>, <A href="http://w3.org//TR/2002/WD-owl-absyn-20020729/">Abstract Syntax</A> and <A href="http://w3.org//TR/2002/WD-owl-ref-20020729/">Language Reference</A> describe the OWL Web Ontology Language 1.0 and its subset OWL Lite. Automated tools can use common sets of terms called ontologies to power services such as more accurate Web search, intelligent software agents, and knowledge management. OWL is used to publish and share ontologies on the Web. Read about the W3C <A href="http://w3.org//2001/sw/">Semantic Web Activity</A>. [<A href="http://www.iaslash.org/">ia/ - news for information architects</A>] <BLOCKQUOTE dir=ltr style="MARGIN-RIGHT: 0px"> » Interesting stuff.  I shall have to see how this relates to Topic Mapping and XTM.  My guess is that OWL is a layer above XTM and is used for providing the contours of the topic map.  </BLOCKQUOTE> http://matt.blogs.it/entries/00000270.html XFML bolsters the semantic web Tue, 15 Oct 2002 19:43:15 +0100 <A href="http://www.highcontext.com/blarchive/2002_10_15.html">Semantic Islands</A>. From :<A href="http://www.steptwo.com.au/columntwo/archives/000309.html">Column Two: Death of keywords</A>: <BLOCKQUOTE>To me, this really highlights the challenges (futility?) of the so-called "semantic web", where everything describes itself, cross-linking happens automatically and accurately, and search engines only return useful results... If we can't get even simple keywords tags to work in practice, what hope is there for RDF, and the rest?</BLOCKQUOTE> My own opinion is that any acitivity or tool that requires consistent, similar, behaviors across the entire Web (such as accurate keywording of web pages) will not happen. However, that doesn't mean the keyword metatag is dead. It can still be an effective tool for a collection of content whose authors/owners are willing to invest time and effort into for accurate searching and indexing. The Web might evolve into small, organized, clusters of content that create semantic islands in a chaotic sea. [<A href="http://www.highcontext.com/">High Context</A>] <BLOCKQUOTE dir=ltr style="MARGIN-RIGHT: 0px"> » Interesting.  One of the ideas behind "!xfml" is to leverage the attempts made by others to index content without requiring a high degree of consistency between resources. Specifically you can attach keywords to page (urls) in an XFML document and the <connect> your keywords to the keywords in other documents.</BLOCKQUOTE> http://matt.blogs.it/entries/00000728.html Semantic blogging Mon, 24 Feb 2003 14:03:46 +0000 <BLOCKQUOTE dir=ltr style="MARGIN-RIGHT: 0px"> <A href="http://www.myelin.co.nz/post/2003/2/24/#200302241">Semantic blogging tool?</A>. Danny Ayers is <A href="http://dannyayers.com/ideagraph-blog/archives/000850.html">working on a 'semantic weblog server'</A>. It'll be interesting to see how this turns out. The semantic web needs working apps, so the rest of us can get a clue re what it's all about. [<A href="http://www.myelin.co.nz/post/">Second p0st</A>]</BLOCKQUOTE> This looks interesting.  http://matt.blogs.it/entries/00001188.html Exploring topics in RSS2.0 Sat, 05 Apr 2003 10:38:33 +0100 I've been doing some thinking about how to encode topic information into RSS2.0 feeds.  As a simple test of the Radio callback facility I have <A href="http://matt.blogs.it/rss.xml">implemented</A> a very simplistic protocol.  Within each <item> is a tag <BLOCKQUOTE dir=ltr style="MARGIN-RIGHT: 0px"> <topic id="topic_id" type="topic-type" source="url">topic name</topic></BLOCKQUOTE> for each topic associated with the item (post).  A concrete example (using the rsstopics namespace): <BLOCKQUOTE dir=ltr style="MARGIN-RIGHT: 0px"> <rsstopics:topic rsstopics:id="the_state" rsstopics:source="http://matt.blogs.it/topics/topicsT.html#the_state" rsstopics:type="generic">the state</rsstopics:topic></BLOCKQUOTE> Whilst this does have the advantage that it's simple and direct it's also a bit silly to invent a new format for topic information when we have two standard culprits available already: <UL dir=ltr> <LI> <DIV><A href="http://www.w3.org/TR/rdf-primer/">Resource Description Framework (RDF)</A> </DIV> <LI> <DIV><A href="http://www.topicmaps.org/xtm/1.0/">XML Topic Maps (XTM)</A></DIV></LI></UL> RDF is a general format for describing resources.  A resource in RDF terms is anything which can be uniquely identified by a URI.  An RDF statement (utilizing <A href="http://dublincore.org/">Dublin Core</A> metadata) that asserts me as the owner of my weblog might look something like: <BLOCKQUOTE dir=ltr style="MARGIN-RIGHT: 0px"> <rdf:Description rdf:about="http://matt.blogs.it">     <dc:Creator>Matt Mower</dc:Creator> </rdf:Description></BLOCKQUOTE> If you cut away the syntactic fluff what this says is: <BLOCKQUOTE dir=ltr style="MARGIN-RIGHT: 0px"> Matt Mower is the Creator of <A href="http://matt.blogs.it/">http://matt.blogs.it</A></BLOCKQUOTE> Referring back to the problem at hand, describing what a post (expressed as an RSS item) is about we could come up with something like: <BLOCKQUOTE dir=ltr style="MARGIN-RIGHT: 0px"> <item rdf:about="permalink">     <topic id="topic_id" type="topic-type" source="url">topic name</topic> </item></BLOCKQUOTE> Which is more or less exactly where we started -- using RDF hasn't altered the solution but it has added some framework around it (in this case adding rdf:about to signal the presence of RDF data within the item).  However we can go a step further.  A useful <A href="http://www.xml.com/pub/a/2001/05/02/semanticwebsite.html">article</A> by <A href="http://www.xml.com/pub/au/74">Eric van der Vlist</A> discusses this very subject and refers to the RSS1.0 taxonomy module. Somewhat counter to what you would expect <A href="http://backend.userland.com/rss">RSS2.0</A> does not follow on from <A href="http://www.purl.org/rss/1.0/">RSS1.0</A>, nor does RSS1.0 follow on from the popular RSS0.9x formats.  RSS1.0 is, depending upon your point of view, a step forward or an aberation.  RSS1.0 uses a modular set of RDF based tags to describe items in the RSS feed.  One such module is the <A href="http://www.purl.org/rss/1.0/modules/taxonomy/">Taxonomy module</A> which is intended to allow classification of RSS channels & items. Using the taxonomy module you create something like: <BLOCKQUOTE dir=ltr style="MARGIN-RIGHT: 0px"> <item rdf:about="permalink">     <taxo:topics>         <rdf:Bag>             <rdf:li resource="topic-uri-1"/>             <rdf:li resource="topic-uri-2"/>         </rdf:Bag>     </taxo:topics> </item></BLOCKQUOTE> Here the <topics> element contains a list (using the RDF defined Bag - or unorderer list - container element) of resources indicating topics that describe the item.  Each resource then has a <topic> element that describes the topic.  It  might look something like: <BLOCKQUOTE dir=ltr style="MARGIN-RIGHT: 0px"> <taxo:topic rdf:about="http://matt.blogs.it/topics/topicsT.html#the_state">     <taxo:link>http://matt.blogs.it/topics/topicsT.html#the_state<taxo:link>     <rsstopics:type>generic</rsstopics:type>     <dc:title>The State</dc:title> </taxo:topic></BLOCKQUOTE> Although it's a jumble of RDF, the RSS1.0 taxonomy module, Dublic Core, and, a custom rsstopics schema this says exactly the same thing as the original: <BLOCKQUOTE dir=ltr style="MARGIN-RIGHT: 0px"> <topic id="topic_id" type="topic-type" source="url">topic name</topic></BLOCKQUOTE> But do we have to deal with such an ugly mess?  Perhaps not.  Our original choices included the XML Topic Maps format.  This is a complete specification for exchanging topic information.  An example of a topic in XTM format might look something like: <BLOCKQUOTE dir=ltr style="MARGIN-RIGHT: 0px"> <topic id="the_state">     <instanceOf>         <topicRef xlink:href="http://www.purl.org/rss-topics/rss-topics#generic">     </instanceOf>     <baseName>         <baseNameString>The State</baseNameString>     <occurence id="the-state-item">         <instanceOf>             <topicRef xlink:href="http://www.purl.org/rss-topics/rss-topics#story">         </instanceOf>         <resourceRef xlink:href="<permalink-uri>">     </occurence> </topic></BLOCKQUOTE> Again this encodes the same information, using a standard format and only one required namespace (that of XTM itself).  A URI such as <A href="http://www.purl.org/rss-topics/rss-topics#generic">http://www.purl.org/rss-topics/rss-topics#generic</A> points at a topic in another map (in this case a topic describing the topic-type generic). The use of XTM comes with a number of advantages with the main one being that there are an increasing number of tools available to process & manipulate it (for example, see <A href="http://www.topicmap.com/topicmap/tools.html">topicmap.com</A>).  However there also a number of problems with this representation when you attempt to embed it within another XML format such as RSS. <UL dir=ltr> <LI> <DIV>It's not clear whether an XTM fragment such as this is valid when used in this way</DIV> <LI> <DIV>Each time a topic is used we will be duplicating it's details, bloating the markup & potentially creating invalid entries</DIV> <LI> <DIV>The <occurence> relation within the <topic> element is technically redundant.  The enclosing <item> indicates the occurrence.  </DIV></LI></UL> One way to avoid these problems would be to embed the topics within the RSS <channel> definition and refer to them from each <item>.  However we still need a way to refer to the topic and XTM doesn't provide this.  If we had a good way to reference topics then we could either embed mini topic map within the RSS file, or just have the <topicmap> in an external file and point to it.  What could we use?  One possibility is RDF. Using a combination of RDF and XTM would mean something like: <BLOCKQUOTE dir=ltr style="MARGIN-RIGHT: 0px"> <item rdf:about="<permalink-uri>">     <rsstopics:topic>http://www.example.org/myTopicMap.xtm#topic-id</rsstopics:topic>     </item></BLOCKQUOTE> or <BLOCKQUOTE dir=ltr style="MARGIN-RIGHT: 0px"> <item rdf:about="<permalink-uri>">     <rsstopics:topic>#topic-id</rsstopics:topic>   </item></BLOCKQUOTE> In this example the item now refers to an XTM defined topic either elsewhere in the RSS feed (contained within a valid <topicmap> element) or within an external topic map.  The referenced <topic> element can further describe the topic (names, types and so on) using all the expressiveness of XTM.  It's also efficient since there is no duplicated information within the feed. I have described approaches using RDF, XTM and a hybrid of the two.  Each has advantages and disadvantages although I believe the hybrid makes the best use of both formats. I'd welcome comments and or opinions from interested parties. http://matt.blogs.it/entries/00001355.html Cast off your discreteness Thu, 01 May 2003 18:58:39 +0100 <BLOCKQUOTE dir=ltr style="MARGIN-RIGHT: 0px"> <A href="http://www.weblogsky.com#200198758">A note about blogs</A>. I just posted this in a private space, but thought it was worth reposting here: Some of us are seeing weblogs as an early step in the evolution of the web (or, some say, a revolution in the way the web is used), and the general label for the stuff we're talking about is "social software." Social software supports group forming, an activity that wasn't necessarily in the heads of the folks who created the first blog systems as simple content management, emphasizing individual publication. Blogs are evolving, though, as nodes in social networks, and bloggers are drawn to group-forming activities and software developments that emphasize the connections as well as the nodes. It's possible to see blogs as a bunch of discrete publications that order random posts in reverse chronological order, but you get away from that pretty quickly when you get into the space and se what people are actually doing with their weblogs. <A href="http://www.quicktopic.com/21/H/tVuJNMqRSGUaT">Discuss A note about blogs</A> [<A href="http://www.weblogsky.com">weblogsky</A>]</BLOCKQUOTE> Interesting.  The work that <A href="http://paolo.evectors.it/">Paolo</A> and I are doing is about elevating blogs from their discreteness and into a world of connections made through topics! http://matt.blogs.it/entries/00001437.html Presenting People Centred Knowledge Management at the CiG Thu, 29 Apr 2004 18:16:24 +0100 On Tuesday night I did my first proper speakers gig, giving a 20 minutes presentation of People Centred Knowledge Management (PCKM) to members of the <a href="http://www.cityinformation.org.uk">City Information Group</a> (I'll link to their event page when it's been updated). I had a great time doing the event and I've had some positive feedback - I hope everyone there got something out of it. My thanks to the <a href="http://www.cityinformation.org.uk/Committee/committee.htm">organizers</a> Jackie, Genevieve and Nick. <ul> <li><a href="http://matt.blogs.it/gems/presentations/CIG_apr_2004/CIG%20Presentation.htm">Web</a> version of the presentatin. (Should work in all browsers, but you know PowerPoint) </li> <li><a href="http://matt.blogs.it/gems/presentations/CIG_apr_2004/CIG%20Presentation_export.pdf">PDF</a> version of the presentation here. (447K)</li> <li><a href="http://matt.blogs.it/gems/maps/cig_apr_2004/Speakers%20notes.html">Speakers notes</a> (This will give you a better idea of what I said).</li> </ul> I'd like to express my thanks to <a href="http://paolo.evectors.it/">Paolo</a>, <a href="http://www.dijest.com/aka/">Phil</a>, <a href="http://www.hyperorg.com/blogger/">David</a>, and <a href="http://chocnvodka.blogware.com/blog">Suw</a> who were all invaluable in helping me to get prepared. I think it really paid off. Update: It occurs to me that really you don't get very much from my slides. The presentation was a lot about me talking, waving my arms and hopping up and down. You don't get that from PowerPoint. Next time I'd like to be able to webcast the presentation. Anyone have any advice about that sort of thing? http://matt.blogs.it/entries/00002140.html A cracking good read Fri, 17 Jun 2005 23:53:52 +0100 Today I picked up a copy of <a href="http://www.amazon.co.uk/exec/obidos/ASIN/0262531410/qid=1119048812/sr=8-1/ref=sr_8_xs_ap_i1_xgl/202-9212189-3436646">Statistical Language Learning</a> by Eugene Charniak. It seems like quite a thorough treatment of natural language processing using hidden markov models and probabilistic context free grammars. I'm also half way through Thomas Passin's <a href="http://www.amazon.co.uk/exec/obidos/ASIN/1932394206/qid%3D1119048989/202-9212189-3436646">Explorers Guide to the Semantic Web</a> which is a very useful guide to the current state of the art in RDF w.r.t. agents, searching, logic, and ontology. I'm especially motivated by the relevance to the cognitive psychology reading I did on judgement, reasoning, and decision making this year. Off the back of that I'm tinkering with a simple agent programming environment, written in Ruby, which I might release if it amounts to anything interesting. http://matt.blogs.it/entries/00002673.html