Curiouser and Curiouser! on k-collector http://matt.blogs.it/ RSS feed for topic k-collector Copyright 2006 Matt Mower Squib/0.1 self@mattmower.com self@mattmower.com en-gb Testing the k-collector client Wed, 07 May 2003 12:59:50 +0100 <P>Testing the new <FONT color=maroon>k-collector</FONT> client for Radio.&nbsp; This is a cut down version of liveTopics that works with the <A href="http://k-collector.evectors.it/">k-collector server</A>.</P> <P>The k-collector client, unlike liveTopics, is dynamic and based upon a shared topic roll held on the k-collector server.&nbsp; Compared to liveTopics k-collector is stripped down, however it is also smaller, lighter and more stable.</P> <P>I myself am switching over from using liveTopics to using k-collector so that we can test the new server properly.&nbsp; I'm not sure what will happen to the liveTopics client at this point.</P> http://matt.blogs.it/entries/00001454.html More about k-collector Wed, 07 May 2003 15:08:52 +0100 <P>A little bit of background about the k-collector client.</P> <P>k-collector client for Radio Userland is really a totally stripped down version of liveTopics.&nbsp; Out go tables of contents, local databases of topics, lots of macros, XFML, XTM and a raft of other stuff i'm like.</P> <P>What you end up with is a simple client designed to bootstrap itself from an online cloud of topics.&nbsp; For an example of such a cloud is <A href="http://k-collector.evectors.it/itentdirectory/topicRoll.opml?dir=8">here</A>.&nbsp; Topics created on the local system are exported via the RSS feed using <A href="http://www.purl.org/NET/ENT/1.0/">ENT</A> and from there appear on the k-collector server.&nbsp; Each k-collector client regularly polls the cloud looking for new topics and makes them available locally.</P> <P>It's a nice, simple, dynamic system for publishing using shared topics.</P> <P>What k-collector doesn't have most people probably won't miss (especially if they are using a k-collector server) and as a result k-collector is much smaller, lighter and more stable than liveTopics.&nbsp; liveTopics is a complicated application and the combination of Radio & Usertalk don't really support complexity very well.&nbsp; k-collector should suffer from far less problems than it's bigger counterpart.</P> <P>More later.</P> http://matt.blogs.it/entries/00001456.html I'm lazy but that's not the reason Mon, 12 May 2003 12:42:47 +0100 I haven't been posting much over the last week.&nbsp; This is largely because i've been hard at work developing the k-collector client and switching over from liveTopics.&nbsp; That switch is very nearly complete and hopefully from the end of today Paolo and I will be working from the same cloud of topics hosted on our k-collector <A href="http://k-collector.evectors.com/">demo site</A>. http://matt.blogs.it/entries/00001458.html k-collector goes live Tue, 13 May 2003 13:45:29 +0100 <P><A href="http://paolo.evectors.it/">Paolo</A> and I are now subscribed to a single shared cloud (called <A href="http://k-collector.evectors.it/itentdirectory/topicRoll.opml?dir=140">WWWW</A>) of topics using the <A href="http://k-collector.evectors.it/itentdirectory/home?dir=140">k-collector server</A> and client for Radio.</P> <P>This means that our posts will be aggregated together by k-collector on the basis of the topics we use.&nbsp; The demo interface shows a simple hierarchical view but we have lots of other things planned.</P> <P>Another poweful feature of this setup is the shared topic roll.&nbsp; Because we are both subscribed to the WWWW topic roll we use the same topics and any topics we create are automatically made available to other subscribers.</P> <P>For example I am creating a topic 'Paolo Valdemarin' to attach to this post.&nbsp; In a little while Paolo's client will automatically have this new topic available for those moments when he wishes to talk about himself!</P> http://matt.blogs.it/entries/00001459.html Communal topics and super-blogs Tue, 13 May 2003 18:04:02 +0100 <P>Something we've just got working today (okay <A href="http://simone.blogs.it/english/">Simone</A> got it working, but I did a valuable job cheering him on) is one of the final pieces in the k-collector puzzle.&nbsp; The server now sends the client a URL corresponding to each topic in the cloud.&nbsp; This enables the client to create on-the-fly links back to the server referring browsers to a live page about the topic.</P> <P>liveTopics had the concept of the table of contents, a local directory of your topics and what posts they related to.&nbsp; All very well in it's way but a far cry from the super-blogs I <A href="http://matt.blogs.it/2002/07/10.html#a193">wrote about</A> last July.&nbsp; Specifically it wasn't live (you couldn't do anything with that data) and it wasn't shared, MY topics MY posts, bleugh!</P> <P>If you click a topic name on my weblog <EM>now </EM>you don't get a local page but, instead,&nbsp;the dynamic k-collector page for that topic.&nbsp; At the moment this is an aggregation of all the posts about that topic from anyone subscribing to the cloud.&nbsp; Soon it will be much more.&nbsp; k-collector has become my super blog and pretty soon we aim to have many more people subscribing.&nbsp; We also hope to see other people creating clouds and boostrapping their own topic communities.</P> http://matt.blogs.it/entries/00001463.html Wanna know how it works? Tue, 13 May 2003 20:38:08 +0100 <BLOCKQUOTE dir=ltr style="MARGIN-RIGHT: 0px"> <P><A href="http://k-collector.evectors.it/">K-collector</A>. Matt has already written about k-collector today <A href="http://matt.blogs.it/2003/05/13.html#a916">here</A> and <A href="http://matt.blogs.it/2003/05/13.html#a919">here</A>. What can I add? Maybe a little drawing? <IMG src="http://paolo.evectors.it/myImages/entWorkflow3.jpg"> [<A href="http://paolo.evectors.it/">Paolo Valdemarin: Paolo's Weblog</A>]</P></BLOCKQUOTE> <P>Yeah!&nbsp; That's how it works.</P> http://matt.blogs.it/entries/00001465.html Call for help: Cross domain scripting Fri, 06 Jun 2003 16:19:34 +0100 <P>We've run into a problem implementing the <A href="http://www.movabletype.org/">MovableType</A> client for <A href="http://wwww.evectors.it/">k-collector</A>:</P> <P>Our approach has been to bootstrap development by reusing a lot of the Radio client code (moved onto the k-collector server) and having the MT client call that as far as possible.&nbsp; We made some good progress this way however we've hit a snag.&nbsp; When the server side code needs to send information back to MovableType we hit the 'ol Javascript <A href="http://www.mozilla.org/projects/security/components/same-origin.html">same origin policy</A> problem.</P> <P>What we need is a way for the pages serverd by MovableType to allow the pages from k-collector access to some of their properties but without having to use signed pages (which appears to be a minefield).&nbsp; Since the one should be allowed to trust the other I am hoping there is a way.&nbsp; Is there?</P> http://matt.blogs.it/entries/00001471.html Progress on k-collector for MovableType Wed, 11 Jun 2003 19:38:17 +0100 With some much needed help from <A href="http://www.crystalflame.net/">Richard</A> work on the <A href="http://wwww.evectors.it/">k-collector</A> client for <A href="http://www.movabletype.org/">MovableType</A> is proceeding.&nbsp; I won't say smoothly, my <A href="http://www.perl.org/">perl</A> is too ancient, and rusty, and MovableType too complex for that, but it is proceeding.&nbsp; We're hoping to invite the first beta testers to get involved sometime next week. http://matt.blogs.it/entries/00001480.html k-collector DMOZ style Thu, 12 Jun 2003 07:34:04 +0100 <P>Richard Soderberg is <A href="http://www.crystalflame.net/archives/000075.html">wondering</A> whether there is utility in linking k-collector with a <A href="http://www.dmoz.org/">DMOZ</A> style directory.&nbsp; The principle is that DMOZ editors have already come up with a wide-ranging category hierarchy so let people re-use it.</P> <P>I can see this being of benefit to a public k-collector server such as our beta but perhaps of far more limited utility for a company or team using k-collector internally.</P> <P>What do people think?</P> http://matt.blogs.it/entries/00001483.html K-Collector white paper Tue, 12 Aug 2003 19:59:12 +0100 I've spent a big chunk of today writing an email to a suspect in the US.&nbsp; It occurs to me that, as an email, it fails to be concise yet precise.&nbsp; However it would make the basis for a K-Collector white paper and that's a good&nbsp;thing since I have been stuck on that task for a while now. http://matt.blogs.it/entries/00001541.html K-Collector & Bayesian filtering Sun, 17 Aug 2003 18:59:05 +0100 <BLOCKQUOTE> <P><A href="http://www.decafbad.com/blog/geek/bayes_agg_one.phtml">Issues in using SpamBayes to filter news items</A>. </P> <P>Despite a reading <A href="http://www.srijith.net/trinetre/archives/2003/08/11/index.shtml#000373">an entry by Srijith</A> discussing Bayes-based classification as unsuitable for use in news aggregators, I tied <A href="http://www.spambayes.org">SpamBayes</A> into my <A href="http://www.decafbad.com/viewcvs.cgi/dbagg/">homebrew news aggregator</A> and have been trying it out this week. I know I’ve been <A href="http://www.decafbad.com/blog/geek/syndicated_whuffie.phtml">talking about it</A> for awhile, but procrastination and being busy all round kept me from getting to it. Funny thing is, when I finally got a chance to really check things out, the integration was a snap. I’d anticipated a bit of work, but was pleasantly surprised. I doubt that any other aggregator written in Python would have a hard time with it.</P> <P>If, that is, anyone else wants to do it. I already knew it wasn’t <A href="http://diveintomark.org/archives/2003/07/07/linkdumps_are_like_sex">magic pixy dust</A> but I figured it might be worth a try. I will be eating my dogfood for awhile with this, but I’m thinking already that what’s good for spam might not be so good for news aggregators.</P> <P>Srijith’s <A href="http://www.srijith.net/trinetre/archives/2003/08/11/index.shtml#000373">post</A> mentions some snags in ignoring some of the semantics of a news item, such as whether a word appears in the item’s title or information about the item’s source. I don’t think that this completely applies to how I’m doing classification, since SpamBayes appears to differentiate between words found in email headers and the body itself. When I feed an item to SpamBayes for training and scoring, I represent it as something like an email message, with headers like date, subject, from, and an “X-Link” header for the link. However, even with this, I think Srijith’s got a point when he writes that this method will miss a lot of available clues for classification.</P> <P>Unlike Srijith’s examples, though, I’m not trying to train my aggregator to sift entries into any specific categories. So far, I’ve been trying to get it to discriminate between what I really want to read, and what I’m not so interested in. So, I figured that something which can learn the difference between spam and normal email could help. But, although it’s early, I’m noticing a few things about the results and I’ve had a few things occur to me.</P> <P>See, in the case of ham vs spam, I really want all the ham and none of the spam. A method to differentiate between these two should be optimized toward one answer or the other. SpamBayes offers “I don’t know” as a third answer, but it’s not geared toward anything else in-between. However, in measuring something like “interest“, inbetween answers are useful. I want all of the interesting stuff, some of the sort-of interesting stuff, and a little of the rest.</P> <P>This is also a problem for me in deciding to what I should give a thumbs up and what gets the thumbs down. Even though I’ve subscribed to a little over 300 feeds, every item from each of them is somewhat interesting to me. I wouldn’t have subscribed to the feed if there wasn’t anything of interest there, so I’ve already biased the content of what I receive. Some items are more interesting than others, but the difference between them is nowhere near the difference of wanted ham vs unsolicited spam. So, I find myself giving the nod to lots of items, but only turning down a few. SpamBayes would like equal examples of both, if possible.</P> <P>I’ll still be playing with this for awhile, but I need to look around at other machine learning tech. I’m just hacking around, but the important thing is to try to understand the algorithms better and know how they work and why. Bayes is in vogue right now, but as Mark Pilgrim intimated, it’s not magic. It’s just “advanced” :)</P> <P>In the immortal words of <A href="http://www.spidereyeballs.com/os6/set3/small_os6_d3_3596_sm.html">Mark Jason Dominus</A>: “You can’t just make shit up and expect the computer to know what you mean, retardo!” </P> <P>[<A href="http://www.decafbad.com/blog">0xDECAFBAD</A>]</P></BLOCKQUOTE> <P>Interesting.&nbsp; Within <A href="http://w4.evectors.it/">k-collector</A> we already have a method for selecting a level of interest more granular than the feed and that is the topic (and, soon, groups of related topics).</P> <P>This allows you to say <EM><FONT color=maroon>I'm interested in Java</FONT></EM> rather than <EM><FONT color=maroon>I want to read these 200 blogs where they talk about Java sometimes</FONT></EM>.&nbsp; Then within this view you could start to say <EM><FONT color=maroon>well, so and so&nbsp;is more interesting than Matt on this topic</FONT></EM>.&nbsp; But again, you are only dealing with the topic at hand.&nbsp; You might still think I'm more interesting about something else even if I struggle for an example!</P> <P>We're already looking at interesting things we can do with this approach, maybe Bayesian filtering is something we should be thinking about.</P> http://matt.blogs.it/entries/00001544.html When aggregators attack Mon, 18 Aug 2003 14:28:23 +0100 <BLOCKQUOTE dir=ltr style="MARGIN-RIGHT: 0px"> <P><A href="http://scriptingnews.userland.com/2003/08/18#wiredOnAggregators">Comments here</A> on Wired's <A href="http://www.wired.com/news/infostructure/0,1377,60053,00.html">article</A> on news aggregators. [<A href="http://www.scripting.com/">Scripting News</A>]</P></BLOCKQUOTE> <P>The title of the article &nbsp;<FONT color=darkblue>Aggregators Attack Info Overload</FONT>&nbsp; suggested to me a slightly more in depth piece looking at the value of using RSS and aggregators to communicate and share.&nbsp; In this, I think it fell short.</P> <P>I'm also disappointed because it meant that <A href="http://w4.evectors.it/">K-Collector</A> didn't get a mention.&nbsp; We may be the new kid on the block but with RSS+ENT I think we're doing something really interesting in this space.</P> http://matt.blogs.it/entries/00001552.html Bayes makes topics Tue, 19 Aug 2003 19:57:08 +0100 <P>Doing a bit of digging into Bayesian filtering.&nbsp; Although it may have other uses in K-Collector later on, my initial thoughts are using a Bayesian filter to automatically suggest topics for weblog posts via K-Collector client.&nbsp; If it's good enough we might even be able to skip the user having to approve some topics.</P> <P>At the moment the client is using a simple keyword stemmer which is effective, up to&nbsp;a point, but suggests a lot of false positives.&nbsp;&nbsp; I'm hoping that a trained Bayesian classifier will do a lot better.&nbsp; Of course this raises the issue of how it gets trained but that is a bridge of a very different colour.</P> <P>Some resources I have come across:</P> <UL> <LI><A href="http://www.paulgraham.com/spam.html">A Plan for Spam <Paul Graham></A></LI> <LI><A href="http://sourceforge.net/docman/display_doc.php?docid=13334&group_id=63137">Project POPFile</A></LI></UL> http://matt.blogs.it/entries/00001559.html The whole is greater than the sum of it's parts Wed, 20 Aug 2003 10:01:18 +0100 <BLOCKQUOTE dir=ltr style="MARGIN-RIGHT: 0px"> <P><A href="http://weblog.halmacomber.com/2003_08_17_archive.html#106135017216163658">Doing My Best Not to Scream</A>. </P> <P><A href="http://weblog.halmacomber.com/2003_08_17_archive.html#106126040209696149">Yesterday's posting</A> hit a nerve. (Seems at least three people agree with me!) What might we be able to accomplish on our projects if we put our attention on learning to increase the relatedness of people on our projects rather than studying for the <A href="http://www.pmi.org/">PMI</A> certification exam? Does anyone really think that doing better work breakdown structures will make our projects successful? No one. That's what I thought. How about learning to repair trust between two important team members? Now that would make a difference. Not the role of a project manager, you say? Then who's role is it?</P> <P>It's time we stopped acting like good technical wisdom is what makes for good project management. It doesn't. Likewise, accountability, authority, and responsibility (someone needs to explain the difference between accountability and responsibility for me) don't make a project manager. Let's try care, guidance, attention, listening, and openness. Now we're getting somewhere!</P> <P>I recognize my mood in writing this is somewhat impertinent. Frankly, I'm doing my best not to scream. (It would wake the dogs.) We must shift our conversation about project management from the things we do to the people we do it with. Only when we put people at the center of projects can we have the fantastic environments that projects are for our clients, for us and our team mates, and our companies.<BR>[<A href="http://weblog.halmacomber.com/">Reforming Project Management</A>]</P></BLOCKQUOTE> <P>I missed the original post in my aggregator but agree strongly with what is said here.&nbsp; One of the key things I have learned over the last year or so is that it is people that do valuable things and they don't do them on their own. <P>To take my own example, my productivity has soared since I started collaborating with <A href="http://paolo.evectors.it/">Paolo</A> and <A href="http://simone.blogs.it/">Simone</A> of evectors.&nbsp; It's not our technical knowledge that has delivered this, it's the new interactions that we can provoke & sustain&nbsp;in each other.&nbsp; The whole definitely is greater than the sum of it's parts. <P>We're looking at ways in which <A href="k-collector">K-Collector</A> can help to connect people together in organsiations because we think it's a powerful tool.</P> http://matt.blogs.it/entries/00001567.html Aggregator work Wed, 27 Aug 2003 14:24:51 +0100 <P>We're rewriting part of the K-Collector aggregator using a new design.</P> http://matt.blogs.it/entries/00001593.html Many heads in jars Wed, 27 Aug 2003 20:32:32 +0100 <P>It's an amazing group of <A href="http://k-collector.evectors.it/itentdirectory/wwwwclassification?dir=142">people</A> K-Collector has acquired in just a few short months.&nbsp; I wish I knew half these people or at least who they were!</P> <DIV><SPAN><SPAN>Adam Curry,</SPAN></SPAN><SPAN>Adriaan Tijsseling,</SPAN><SPAN>Adrian Scott,</SPAN><SPAN>Alan Cooper,</SPAN><SPAN>Alf Eaton,</SPAN><SPAN>Alison Fish,</SPAN><SPAN>Andre Durand,</SPAN><SPAN>Barry Diller,</SPAN><SPAN>Ben Hammersley,</SPAN><SPAN>Ben Trott,</SPAN><SPAN>Beppe Caravita,</SPAN><SPAN>Bill Kearney,</SPAN><SPAN>Brewster Kahle,</SPAN><SPAN>Britt Blaser,</SPAN><SPAN>Cesare Lamanna,</SPAN><SPAN>Chris Pirillo,</SPAN><SPAN>Chuck Lawson,</SPAN><SPAN>Clay Shirky,</SPAN><SPAN>Cory Doctorow,</SPAN><SPAN>Dan Brickley,</SPAN><SPAN>Dan Gillmor,</SPAN><SPAN>danah boyd,</SPAN><SPAN>Danny Ayers,</SPAN><SPAN>Danny Goodman,D</SPAN><SPAN>ave Jacobs,</SPAN><SPAN>Dave Snowden,</SPAN><SPAN>Dave Winer,</SPAN><SPAN>David Galbraith,</SPAN><SPAN>David Gurteen,</SPAN><SPAN>David Isenberg,</SPAN><SPAN>David Sifry,</SPAN><SPAN>Dina Mehta,</SPAN><SPAN>Doc Searls,</SPAN><SPAN>Don Hopkins,</SPAN><SPAN>Don Norman,</SPAN><SPAN>Don Park,</SPAN><SPAN>Doug Baron,</SPAN><SPAN>Drummond Reed,D</SPAN><SPAN>uncan Wilcox,E</SPAN><SPAN>dd Dumbill,</SPAN><SPAN>Eric Norlin,</SPAN><SPAN>Eugene Kim,</SPAN><SPAN>Evan Williams,</SPAN><SPAN>Gary Santoro,</SPAN><SPAN>Giuseppe Granieri,</SPAN><SPAN>Greg Elin,</SPAN><SPAN>Harold Gilchrist,</SPAN><SPAN>Howard Dean,</SPAN><SPAN>Howard Rheingold,</SPAN><SPAN>Jason DeFillippo,</SPAN><SPAN>Jason Fried,</SPAN><SPAN>Jason Lefkowitz,</SPAN><SPAN>Jay Feinberg,</SPAN><SPAN>Jim McGee,</SPAN><SPAN>John Bransford,</SPAN><SPAN>John Doerr,</SPAN><SPAN>John Hagel,</SPAN><SPAN>John Norris,</SPAN><SPAN>John Robb,</SPAN><SPAN>John Seely Brown,</SPAN><SPAN>Joi ito,</SPAN><SPAN>Jon Lebkowsky,</SPAN><SPAN>Jon Udell,</SPAN><SPAN>Jonathan Abrams,</SPAN><SPAN>Jonathan Peterson,</SPAN><SPAN>Justin Hall,</SPAN><SPAN>Karlin Lillington,</SPAN><SPAN>Kevin Marks,</SPAN><SPAN>Kevin Werbach,</SPAN><SPAN>Kim Polese,</SPAN><SPAN>Lawrence Lessig,</SPAN><SPAN>Leigh Dodds,</SPAN><SPAN>Libby Miller,</SPAN><SPAN>Lisa Rein,</SPAN><SPAN>Liz Lawley,</SPAN><SPAN>LockerGnome,</SPAN><SPAN>Marc Barrot,</SPAN><SPAN>Mark Carey,</SPAN><SPAN>Mark Pilgrim,</SPAN><SPAN>Matt Webb,</SPAN><SPAN>Meg Hourihan,</SPAN><SPAN>Micah Alpern,</SPAN><SPAN>Michael Gartenberg,</SPAN><SPAN>Mitch Ratcliffe,</SPAN><SPAN>Morten Fredericksen,</SPAN><SPAN>Nat Friedman,</SPAN><SPAN>Neo,</SPAN><SPAN>Nick Gaydos,</SPAN><SPAN>Nikolaj Nyholm,</SPAN><SPAN>Ole Eichhorn,</SPAN><SPAN>Paolo Valdemarin,</SPAN><SPAN>Paul Boutin,</SPAN><SPAN>Peter Kaminsky,</SPAN><SPAN>Phil Pearson,</SPAN><SPAN>Phil Wolff,</SPAN><SPAN>Pierre Omidyar,</SPAN><SPAN>Ray Ozzie,</SPAN><SPAN>Reid Hoffman,</SPAN><SPAN>Richard MacManus,</SPAN><SPAN>Rob Adams,</SPAN><SPAN>Robb Beal,</SPAN><SPAN>Robert Scoble,</SPAN><SPAN>Ross Mayfield,</SPAN><SPAN>Russell Beattie,</SPAN><SPAN>Sam Ruby,</SPAN><SPAN>Scott Adams,</SPAN><SPAN>Scott Adans,</SPAN><SPAN>Scott Heiferman,</SPAN><SPAN>Scott Johnson,</SPAN><SPAN>Scott Mace,</SPAN><SPAN>Seb Paquet,</SPAN><SPAN>Shelly Powers,</SPAN><SPAN>Simone Bettini,</SPAN><SPAN>Sjoerd Visscher,</SPAN><SPAN>Steve Yost,</SPAN><SPAN>Thomas Burg,</SPAN><SPAN>Thomas Mygdal-Madsen,</SPAN><SPAN>Tim O'Reilly,</SPAN><SPAN>Ward Cunningham,</SPAN><SPAN>Will Wright,</SPAN><SPAN>Xeni Jardin,</SPAN><SPAN>Zack Lynch</SPAN></DIV> http://matt.blogs.it/entries/00001599.html A journey with Phil Tue, 23 Sep 2003 20:33:41 +0100 So, I don't have the patience to do a complete re-write but here is a recap of some of the highlights of my <a href="http://dijest.com/aka/2003/09/22.html#a2623">conversation with Phil</a>.<br> <br> Obviously, since we were using it, we talked about Skype (i'm pronouncing it 'sky'-'p'). I guess I'm both impressed and frustrated in equal measure. Most of the time (we were talking for hours, god knows what state Phil was in the next day) the quality was good, but it did crap out on us quite a bit. It definitely needs a cell-phone style signal meter. We even mused that, since Skype knows your address it could tap into the <a href="http://www.internettrafficreport.com/main.htm">Internet Weather Report</a> to tell you what sort of call to expect!<br> <br> We talked about the possible applications for a good, free, VOIP client and there are many. One in particular appeals to me and that is seeing Skype support bundled with the software I buy. What I want is that when I need to contact a vendor I can press a <span style="font-style: italic;">Skype me</span> button. This lets them know that a CSA or Techie should get in touch with me. It's convenient for me because it means that I will be at the computer when they call. The win for the vendor is that they only need to call when they see me online and the cost is significantly cheaper than how they do this today. (As an example, doing webcallback via Netcall means the vendor has to pay for 2 PSTN calls. One from the Netcall server to the CSA and one from the Netcall server to the customer. Then they have to pay Netcall to manage it all as well. Gets expensive).<br> <br> Phil and I are both lefties so we rapped about the war, Bush and the Dean campaign. It's incredibly for a guy that seemed such an outside 12 months ago that if you search for Howard on Google, <a href="http://www.deanforamerica.com/">Dean for America</a> is no. #2. What is particularly impressive is the way the Dean campaign have leveraged their digital savvy into on the ground support. Would that any UK politician had the same nouse. I can't imagine being so impressed with any MP I have come across that i'd actually go talk about them. <span style="background-color: rgb(255, 255, 255); color: rgb(153, 0, 0);">I would love to feel differently though!<br> <span style="color: rgb(51, 0, 51);"><br> I gave Phil a quick head's up on the work we are doing with <a href="http://w4.evectors.it/">K-Collector/W4</a>, he hadn't seen it since <a href="http://blogtalk.net/">BlogTalk</a> (it seems so long ago now...) and we've put a lot of effort in since then.&nbsp; We also <a href="http://dijest.com/aka/2003/09/22.html#a2623">talked about what's needed to get corporates into blogging</a>.<br> <br> Oh he put me onto <a href="http://www.amazon.co.uk/exec/obidos/ASIN/0451459156/ref=sr_aps_books_1_2/026-1765779-1917222">Ruled Brittania</a> (by Harry Turtledove) which is an alternate reality novel set in an England conquered by Spain.&nbsp; I'm a complete sucker for this stuff so that went straight onto my <a href="http://www.amazon.co.uk/exec/obidos/registry/2VBDDIPM7ZWQP/ref=wl_s_3/026-8728199-5122817">wishlist</a>.&nbsp; In return I suggested <a href="http://www.amazon.co.uk/exec/obidos/ASIN/0099263815/ref=sr_aps_books_1_1/026-1765779-1917222">Fatherland</a> (by Robert Harris) a detective thriller set in a victorious post WW II Germany and <a href="http://www.amazon.co.uk/exec/obidos/ASIN/014017172X/ref=sr_aps_books_1_2/026-1765779-1917222">The Man in the High Castle</a> (by Philip K. Dick and personal favourite of mine) set in an America split down the middle by Germany and Japan.<br> <br> There was lots more but I'm running out of steam...<br> </span></span> http://matt.blogs.it/entries/00001646.html K-Collector top 10 since last night Wed, 08 Oct 2003 09:28:18 +0100 Last night I enabled some new code in K-Collector to track visitors to each topic. In combination with other data we are collecting in the system it should allow us to do some neat tricks, but for now it tells us which topics people are visiting the most. The top 10 in descending order:<br> <br> <a href="http://k-collector.evectors.it/itEntDirectory/w4topicbyid?id=knowledge_management">Knowledge Management</a><br> <a href="http://k-collector.evectors.it/itEntDirectory/w4topicbyid?id=don_park">Don Park</a><br> <a href="http://k-collector.evectors.it/itEntDirectory/w4topicbyid?id=robert_scoble">Robert Scoble</a><br> <a href="http://k-collector.evectors.it/itEntDirectory/w4topicbyid?id=bloggercon">BloggerCon</a><br> <a href="http://k-collector.evectors.it/itEntDirectory/w4topicbyid?id=joi_ito">Joi Ito</a><br> <a href="http://k-collector.evectors.it/itEntDirectory/w4topicbyid?id=ideas">Ideas</a><br> <a href="http://k-collector.evectors.it/itEntDirectory/w4topicbyid?id=roland_tanglao">Rolang Tanglao</a><br> <a href="http://k-collector.evectors.it/itEntDirectory/w4topicbyid?id=open_standards">Open Standards</a><br> <a href="http://k-collector.evectors.it/itEntDirectory/w4topicbyid?id=laszlo">Laszlo</a><br> <a href="http://k-collector.evectors.it/itEntDirectory/w4topicbyid?id=k-collector">K-Collector</a><br> <br> The BloggerCon topic has been quite interesting since we added most of the feeds from the BloggerCon blogroll. With the new auto-discovery code in place this has allowed it to link in quite a lot of posts and relate them to other topics in the system.<br> <br> Take a look and let us know what you think.<br> http://matt.blogs.it/entries/00001650.html K-Collector as an RSS catalogue Tue, 14 Oct 2003 10:58:05 +0100 <div style="margin-left: 40px;"><a href="http://www.istori.com/log/archives/00000310.html">RSS Feed Catalogs</a>. Wouldn't it be cool if sites that published lots of RSS feeds could also publish a catalog of those feeds ... [<a href="http://www.istori.com/log/">istori/log</a>]<br> </div> <br> We're publishing several hundred <span style="font-style: italic;">topical </span>feeds (in RSS2.0 + <a href="http://matt.blogs.it/specs/ENT/1.0/">ENT</a> format) from <a href="http://w4.evectors.it/">K-Collector</a>.&nbsp; You can navigate the feed structure quite easily.&nbsp; Start at the <a href="http://w4.evectors.it/">top-level</a> and select a category such as <a href="http://k-collector.evectors.it/itentdirectory/wwwwclassification?dir=142">Who</a> then browse the available topics under that category.&nbsp; Pick one like <a href="http://k-collector.evectors.it/itentdirectory/wwwwtopic?dir=361">Don Park</a> or <a href="http://k-collector.evectors.it/itentdirectory/wwwwtopic?dir=468">Phil Wolff</a> and then add the <a href="http://k-collector.evectors.it/itentdirectory/rss2?dir=361">XML feed</a> to your aggregator to read posts concerning that person.<br> http://matt.blogs.it/entries/00001673.html Knowledge Centrifuge Sun, 26 Oct 2003 21:13:00 +0000 As we get closer to release I've found myself explaining what <a href="http://paolo.evectors.it/">Paolo</a> and I are doing with <a href="http://w4.evectors.it/">K-Collector</a>, what the product is about. So far the best explanation I have come up with is:<br> <br> <div style="text-align: left; margin-left: 40px;"><span style="font-weight: bold;">Realtime Knowledge Management</span><br> </div> <br> What I've been trying to do is make the case that knowledge management and document management aren't the same thing. Most knowledge in a company begins life as a granule of information, micro-knowledge. It exists for a time and, if not exploited, likely dies away to be discovered again later if needed.<br> <br> Quite often the inertia that has to be overcome in order to turn such a granule into the sort of document you would load into Livelink, Documentum or some equivalent product is overwhelming compared to the perceived value, at that moment, of the knowledge itself. It is only when the cost of repeated rediscovery begins to bite that someone finally does the decent thing.<br> <br> The corrollary to this is that the benefits of knowing that you know this information are not felt until quite late in it's lifecycle. I would guess that, by the time a lot of information is formally documented, it's probably well on the well to being out of date or irrelevant (is this your intranet?)<br> <br> By contrast the K-Collector approach collects information while it is still fresh and combines it with other related information before presenting to users for them to see if it meets their needs. That which is good can be promoted to a more appropriate place (for example a <a href="http://www.socialtext.com/">Wiki</a>) Things which don't make the cut fade into the background but, crucially, are not lost.<br> <br> I imagine K-Collector as being a kind of <span style="font-style: italic; color: rgb(0, 0, 153);">knowledge centrifuge</span>, spinning together all kinds of different bits of information and separating out the good stuff for you.<br> <br> At the heart of K-Collector and determining what makes 'the good stuff' are topics. Topics act as markers for points of interest around which information can be clustered. The Who, What, Where, When metaphor we have adopted is - we think - a really simple way of considering what is important to us all (although, prompted by <a href="http://www.henshall.com/blog/">Stuart Henshall</a>, I have been wondering about <span style="font-style: italic;">Why</span> as a 5th <span style="font-style: italic;">W</span>).<br> <br> I think of K-Collector as a kind of multi-dimensional database where each topic slices through the available information. We're working on some pretty cool topic related trickery for future versions that will take this idea and make it a lot more powerful.<br> <br> I'm hoping to get back to writing soon and share more about what we are doing. I hope this is useful in the meantime.<br> http://matt.blogs.it/entries/00001679.html K-Collector for MovableType Wed, 29 Oct 2003 22:57:24 +0000 I've spent a fair bit of time today working on the <a href="http://w4.evectors.it/">K-Collector</a> client for <a href="http://www.movabletype.org/">MovableType</a>.<br> <br> We tried implementing a client about 6 months ago but it was very brittle and, ultimately, even I couldn't write instructions for how to install it on a new server. This didn't seem like a good way forward with MT Pro and Typepad around the cornern.<br> <br> Our new approach we're much happier with. It only requires 1 change (adding 1 line) to an existing MT template, 1 perl module (HTML::HeadParser) and the installation of a handful of new files. It does not depend upon the MT database or other plugins. It's pretty lightweight.<br> <br> There is about half a days worth of clean-up required I think and then it should be ready to beta test. I'm hoping to start that on Friday afternoon, interested parties should look out for an announcement with a link to instructions.<br> <br> I'm grateful to <a href="http://www.mplode.com/tima/">Timothy Appnel</a> for his <a href="http://www.oreillynet.com/pub/a/javascript/2003/03/18/movabletype.html?page=1">guide on writing MovableType plugins</a>. It was really helpful when things just weren't working.<br> http://matt.blogs.it/entries/00001689.html K-Collector client for MT update Mon, 03 Nov 2003 10:12:37 +0000 Over the weekend I finished the basic K-Collector client for MovableType.&nbsp; It's still a little rough around the edges, unoptimized and what have you, but it seems to work here.&nbsp;&nbsp; In the first instance we're looking for a handful of MT users who are technically savvy (happy installing perl modules, CGI & don't mind mucking about with their templates and such) to do the testing, then we'll open it out when we are sure we have any problems locked down.<br> <br> If you would like to be part of this first group please <a href="mailto:k-collector@evectors.com">drop us a line</a>.<br> http://matt.blogs.it/entries/00001697.html Learning from collaborative filtering Thu, 06 Nov 2003 16:24:25 +0000 <BLOCKQUOTE dir=ltr style="MARGIN-RIGHT: 0px"> <UL> <LI><A href="http://www.sims.berkeley.edu/resources/collab/">Collaborative Filtering resources </A></LI></UL> <P>[<A href="http://www.decafbad.com/blog">0xDECAFBAD</A>]</P></BLOCKQUOTE> <P>I need to understand how the work being done on collaborative filtering can help us to improve K-Collector.</P> http://matt.blogs.it/entries/00001703.html Learning templates for organising content Mon, 10 Nov 2003 11:14:26 +0000 I've just come across Denham Grey's <a href="http://denham.typepad.com/km/2003/11/information_gat.html">Information Gathering Template</a> post.&nbsp; I'm not sure how I missed it before, I'm subscribed to Denham's feed.&nbsp; Anyhow it's really interesting.<br> <br> What he's done is to publish his informal framework for categorizing information when doing research.&nbsp; It goes like this:<br> <ul> <li>People</li> <li>Places</li> <li>Problems</li> <li>Promises</li> <li>Principles</li> <li>Patterns</li> <li>Products</li> </ul> Please read Denham's post for his excellent notes on each category.&nbsp; Obviously this is very similar to the approach we have taken with <a href="http://k-collector.evectors.it/">K-Collector</a>.<br> <ul> <li>Who</li> <li>Where</li> <li>What<br> </li> <li>When</li> </ul> to which I think it might make sense to add 'Why'.&nbsp; In this case our framework is simpler, conflating: problems, promises, principles, patterns, and, products into what.<br> <br> Although W4 is less expressive than Denham's template we think making it simpler keeps it suitable for general use (Denham is a consultant and far more experienced than most people.)<br> <br> &nbsp;However if we can find a way to layer additional levels of meaning without giving up that simplicity I think we will do it.&nbsp; It makes sense to enable users who can benefit from that extra expressability and for all users to benefit from the new relations we can build among topics.&nbsp; Since K-Collector's architecture allows us to do this it's just a matter of working out the best way and Paolo and I have been talking about this recently.&nbsp; We'll also be sounding out beta testers and customers to see what they think.<br> <br> I'm keen to hear from anyone about their frameworks for collecting & organising content.&nbsp; I do think we have many lessons to learn and want <a href="http://w4.evectors.it/">K-Collector</a> to reflect that learning as much as possible.<br> http://matt.blogs.it/entries/00001713.html What comes after a title? Mon, 17 Nov 2003 15:32:28 +0000 <div style="margin-left: 40px;"><a href="http://radio.weblogs.com/0001011/2003/11/17.html#a5493">I'm using Titles Now</a>. </div> <p style="margin-left: 40px;">I'm trying to put titles on my items now, which should help out my RSS feeds. Hope they work OK. I'm looking for other ways to improve my weblog's usability. Any other good ideas out there? Of course, if I forget to put the titles on, then what's the use? I really hate having to do more work, but I think it's worth it.</p> <div style="margin-left: 40px;"> [<a href="http://radio.weblogs.com/0001011/">The Scobleizer Weblog</a>]<br> </div> <br> How about adding topics and linking your posts to the shared knowledge space that <a href="http://w4.evectors.it/">K-Collector</a> is becoming?<br> http://matt.blogs.it/entries/00001738.html NewsBayes Thu, 20 Nov 2003 16:23:49 +0000 <div style="margin-left: 40px;"><a href="http://weblog.infoworld.com/udell/2003/11/20.html#a851">Working with Bayesian categorizers</a>. </div> <blockquote cite="O'Reilly Network" style="margin-left: 40px;"> There's been some discussion in the blog world about using a Bayesian categorizer to enable a person to discriminate along various interest/non-interest axes. I took a run at this recently and, although my experiments haven't been wildly successful, I want to report them because I think the idea may have merit. [Full story: <a href="http://www.xml.com/pub/a/2003/11/19/udell.html">O'Reilly Network: Working with Bayesian Categorizers</a>] </blockquote> <div style="margin-left: 40px;">This month's O'Reilly Network column was a struggle because categorization itself is a struggle. I remain convinced that the automated classifiers that are doing such a good job beating back the tide of spam will also turn out to be more generally useful. But finding the right synergy between an automated assistant and a human overseer is a subtle and tricky thing. <b>...</b> [<a href="http://weblog.infoworld.com/udell/">Jon's Radio</a>]<br> </div> <br> Great write-up by Jon of his experiments trying to hotwire Bayesian categorizers for auto-classifying blog posts.<br> <br> I think his conclusions demonstrate that there is value in the approach, certainly as Jon notes, for auto-classification of incoming items as in a news aggregator.&nbsp; Of course this is exactly what we are doing with <a href="http://w4.evectors.it/">K-Collector</a> and I have mused for a while about whether we could use a Bayesian classifier to improve how we do that (at the moment we are using some pretty simple keyword analysis).<br> <br> I think one of the biggest obstactles is the training.&nbsp; We have somewhere around 400 topics and many thousands of posts.&nbsp; How hard would it be to train a classifier?&nbsp; How long would it take?&nbsp; Could you sell it?<br> http://matt.blogs.it/entries/00001745.html What comes after changeblogs? Thu, 20 Nov 2003 16:37:16 +0000 <div style="margin-left: 40px;"><a href="http://www.decafbad.com/blog/links/boocjaajai.html">Quick Links</a>. </div> <ul style="margin-left: 40px;"><li><a href="http://jeremy.zawodny.com/blog/archives/001075.html">CVS Commit + Weblog = Changeblog </a></li></ul> <div style="margin-left: 40px;"> [<a href="http://www.decafbad.com/blog">0xDECAFBAD</a>]<br> </div> <br> <a href="http://www.dijest.com/aka/">Phil Wolff</a> and I were <a href="http://matt.blogs.it/2003/07/21.html#a1042">talking</a> about this a little while ago and now <a href="http://www.multiply.org/notebook/">Jason Gessner</a> has it working.&nbsp; That's great!<br> <br> The next stage for me would be to start creating RSS feeds with <a href="http://matt.blogs.it/spects/ENT/1.0/">ENT</a> metadata from those changeblogs. At a very basic level I can imagine a <a href="http://w4.evectors.it/">K-Collector</a> topic corresponding to each project.&nbsp; Then each topic view could aggregated not only what developers were writing in their own blogs but also CVS messages corresponding to the status of the project.<br> <br> But why stop there.&nbsp; Your junit based daily smoke & build test could be generating a similar feed, your project management system, issue database and so on.&nbsp; All these feeds could be flowing into project topics giving you an uptodate and holistic view of what is happening in those projects.<br> <br> Who'd ever be caught on the hop in a project meeting again?<br> <br> <br> http://matt.blogs.it/entries/00001749.html Whose personality do you want today? Fri, 21 Nov 2003 17:47:08 +0000 <a href="http://www.decafbad.com/blog/">l.m.orchard</a> <a href="http://radiocomments.userland.com/comments?u=0107808&p=1243">commented</a> regarding the <a href="http://matt.blogs.it/2003/11/20.html#a1243">using Bayesian analysis on news</a>. In fact, as soon as I saw it I remembered, I had read his <a href="http://www.decafbad.com/blog/geek/bayes_agg_one.html">piece</a> already. It was probably his writing that triggered my initial interest in using a Bayesian classifier in <a href="http://w4.evectors.it/">K-Collector</a>.<br> <br> Re-reading that piece I got an interesting different angle since his approach was to blend a Bayesian classifier with his news aggregator to try and have it prioritize news he would find interesting and not to categorize it by topic. I think this is a much more scalable task, from a K-Collector perspective, than what <a href="http://weblog.infoworld.com/udell/2003/11/20.html#a851">Jon</a> is experimenting with. I think the efforts of training a system-wide recognizer to differentiate between topics would be too much for most users of the product to bear.<br> <br> Our product roadmap for K-Collector already includes allowing users to personalize the system. For example we think that people should be able to say which feeds they think are relevant on different topics. Notice that this is a much very granular relationship since it means that I can say "Matt Mower is a real expert on the topic sock puppets" but that this says nothing about how relevant I am on "dating." or any other topic. Indeed each user might rate the exact same sources differently over a wide range of topics.<br> <br> What might be interesting is if people could "share" and "subscribe to" preference maps. As a new user of the system you might not really know who is relevant on any particular topic. But imagine you worked with <a href="http://www.hyperorg.com/blogger/">David Weinberger</a>, <a href="http://www.dijest.com/aka/">Phil Wolff</a>, or <a href="http://weblog.siliconvalley.com/column/dangillmor/">Dan Gillmor</a>. If you knew them and trusted their judgement you could pick one of their preference maps as a starting point and immediately gain a usseful insight into the data as it is structured by topic. You might even switch between personalities to get more perspective!<br> <br> Thanks to l.m.'s piece I am now wondering also about whether a Bayesian classifier might be more use in helping users to establish their own preference maps about which content is most relevant to them.<br> http://matt.blogs.it/entries/00001757.html Proud to announce K-Collector 1.0 Mon, 24 Nov 2003 15:14:55 +0000 Can it really be a <a href="http://matt.blogs.it/2002/11/20.html#a555">year </a>since that fateful day when <a href="http://paolo.evectors.it/">Paolo</a> and I scribbed on a whiteboard in Ronchi?&nbsp; Can it really be six months since Paolo, <a href="http://simone.blogs.it/">Simone</a> and I gave the <a href="http://paolo.evectors.it/myImages/blogtalk/21.html">first demo</a> of the K-Collector prototype?<br> <br> Time flies when you're having fun and the last year with my new partners at <a href="http://www.evectors.com/">Evectors</a> (Paolo, Simone, Monica and Fabrizio) has gone so quick my head is spinning.&nbsp; Or maybe it's the champagne because today we are announcing the <span style="font-weight: bold;">launch of <a href="http://www.evectors.com/itkcollector">K-Collector 1.0</a></span>!<br> <br> I'm so proud of what we've achieved together in the last 6 months.&nbsp; The product is better than I could ever have imagined it would be, and the best is yet to come!<br> http://matt.blogs.it/entries/00001766.html I am not a vector. I am a free man! Mon, 24 Nov 2003 15:22:16 +0000 We've posted an <A href="http://www.evectors.com/itkcollector/story$num=4&sec=1&data=kcollector">essay</A> which attempts to highlight where we think organisations can benefit from blogging.<BR><%softShadow("http://paolo.evectors.it/stories/vectorWorkers/images/img6.gif")%><BR>It continues our <SPAN style="FONT-WEIGHT: bold; FONT-STYLE: italic">effective organsiations</SPAN> theme and uses what we think is a very strong visual metaphor for understanding how different organisations can apply similar amounts of effort but get wildy different results. http://matt.blogs.it/entries/00001771.html What do you write after a Tonno e Zucchine sandwich? Tue, 02 Dec 2003 19:57:57 +0000 It's been a good trip - I can't believe I fly back tomorrow.&nbsp; Where did the time go?&nbsp; <a href="http://paolo.evectors.it/">Paolo</a> and I have discussed a lot of the strategy behind how we are going to offer <a href="http://w4.evectors.it/">K-Collector</a> to customers. All those questions we've been putting off till last have come due.&nbsp; I think we've got some good answers and we'll be publishing the details soon.&nbsp; I'll also be spending increasing amounts of my time on sales & marketing now which, strangely, I find I am looking forward to!<br> http://matt.blogs.it/entries/00001780.html First steps with Lucene Tue, 02 Dec 2003 20:43:39 +0000 I'm playing with the <a href="http://jakarta.apache.org/lucene/">Jakarta Lucene</a> search engine at the moment as a way of implementing some neat new stuff for <a href="http://w4.evectors.it/">K-Collector</a>.&nbsp; I've got a Java app talking to K-Collector via XML-RPC and indexing posts from our database.&nbsp; I think it took me about 15 minutes from cold to having butchered the sample apps to do what I want.&nbsp; That seems like a good sign to me.&nbsp; While browsing for more information about Lucene I came across <a href="http://www-106.ibm.com/developerworks/library/j-lucene/">an article</a> talking about Lucene and another Jakarta library called <a href="http://jakarta.apache.org/commons/digester.html">Digester</a>.&nbsp; Digester is a neat XML to Java object mapper.&nbsp; I'll be taking a better look at it over the next few days.<br> http://matt.blogs.it/entries/00001782.html Gimme the highlights Fri, 09 Jan 2004 18:52:59 +0000 <p>Today we've released a new feature for <a href="http://w4.evectors.it/">K-Collector</a> which we think makes related topics more useful.</p> <p>Now when browsing a topic the related topics are rendered as checkboxes. When the checkbox corresponding to a specific post is checked, the posts on the current page associated with that topic are highlighted. This makes it very easy to spot posts referencing a specific combination of topics.</p> <p>For example: want to see what people have said about Howard Dean and Meetup? On the page for <a href="http://w4.evectors.it/itentdirectory/topic?topic=howard_dean">Howard Dean</a> check the topic <em>MeetUp</em> and then quickly scroll down to find just those posts. It's a really quick way of finding something in the view.</p> <p>All feedback welcome.</p> http://matt.blogs.it/entries/00001793.html ETCon'04 - We won't be there Fri, 16 Jan 2004 08:57:46 +0000 <blockquote><a href="http://socialsoftware.weblogsinc.com/entry/4114434910144189/">ETECH is coming up....</a>. <A href="http://socialsoftware.weblogsinc.com/entry/4114434910144189/">O'Reilly's Emerging Technology Conference...</A>. <H2>O’Reilly’s Emerging Technology Conference…</H2> <H3>Posted Jan 15, 2004, 6:54 PM ET by Judith Meskill</H3><IMG src="http://conferences.oreillynet.com/images/et2004/etcon_butterfly.gif" align=right> <P><A href="http://conferences.oreillynet.com/etech/"><FONT color=#660000>O’Reilly’s Emerging Technology Conference</FONT></A> — taking place at the Westin Horton Plaza, San Diego, CA, Feb. 9-12, 2004 — will have a <A href="http://conferences.oreillynet.com/pub/w/28/track_social.html"><FONT color=#660000>Social Software track</FONT></A>. This promises to be an excellent event with a broad spectrum of notable speakers that includes (but is certainly not limited to): <A href="http://conferences.oreillynet.com/cs/et2004/view/e_spkr/1686"><FONT color=#660000>Helen Greiner</FONT></A> - iRobot Corp., <A href="http://conferences.oreillynet.com/cs/et2004/view/e_spkr/521"><FONT color=#660000>Cory Doctorow</FONT></A> - EFF, <A href="http://conferences.oreillynet.com/cs/et2004/view/e_spkr/1730"><FONT color=#660000>Lili Cheng</FONT></A> - Microsoft Research, <A href="http://conferences.oreillynet.com/cs/et2004/view/e_spkr/1727"><FONT color=#660000>Gilman Louie</FONT></A> - In-Q-Tel, <A href="http://conferences.oreillynet.com/cs/et2004/view/e_spkr/363"><FONT color=#660000>David Sifry</FONT></A> - Technorati, <A href="http://conferences.oreillynet.com/cs/et2004/view/e_spkr/1703"><FONT color=#660000>Joichi Ito</FONT></A> - Neoteny, <A href="http://conferences.oreillynet.com/cs/et2004/view/e_spkr/1669"><FONT color=#660000>Elizabeth Lawley</FONT></A> - Rochester Institute of Technology, and, of course, <A href="http://conferences.oreillynet.com/cs/et2004/view/e_spkr/416"><FONT color=#660000>Tim O’Reilly</FONT></A> - O’Reilly & Associates. [<A href="http://socialsoftware.weblogsinc.com/">The Social Software Weblog</A>]</P> <P>This is the key event of the year.&nbsp; We're gonna party like is USED to be 1999.</P> <P>I'll be there - sponsored by <A href="http://laszlosystems.com">Laszlo Systems</A> and I'll be giving a :05 minute talk on FOAF and the PeopleAggregator.</P> <P>But clearly the most exciting event will be the field trip to TJ and the House of Mole.&nbsp; Something not to be missed.</P> [<a href="http://blogs.it/0100198/">Marc's Voice</a>]</blockquote> <p>Anyone out there want to sponsor three Europeans with a kick-ass new RSS based collaborative knowledge organisation tool to go to ET'04?</p> http://matt.blogs.it/entries/00001810.html K-Collector @ Wonder Widgets, Inc. Sat, 17 Jan 2004 17:10:37 +0000 <p>We have recently launched a new <a href="http://www.evectors.com/itkcollector/">K-Collector</a> demo site. This is intended to show prospective customers how K-Collector might look (& be used) in their company. The <a href="http://w4.evectors.it/">W4 portal</a> is interesting but too eclectic to really get an idea of how K-Collector is intended to work.</p> <p>The site in question is for our imaginary company <em>Wonder Widgets, Inc.</em> who are purveyors of fine Java and .NET components. We had a bit of fun making up lots of posts ...it's how I imagine soap opera are written... we've already had one person leave the company!</p> <p>So, if you've wondered what K-Collector is <em>really</em> all about, please drop me a line and I will organise a short (10-20 mins) demo for you. I'm happy to do this as I need the practice :-)</p> http://matt.blogs.it/entries/00001822.html K-Collector demo site Sun, 18 Jan 2004 08:57:58 +0000 <p>A quick clarification. I notice that some people read my previous <a href="http://matt.blogs.it/2004/01/17.html#a1294">post</a> and thought the <a href="http://www.evectors.com/itkcollector/">K-Collector</a> link in that post was pointing at the demo site I referrred to.</p> <p>In fact I has linking to the K-Collector page on the <a href="http://www.evectors.com/">Evectors</a> website. This is because the demo site is behind a login at the moment -- we want to make sure that people who use the demo site get the best possible experience. I realise now that this was not at all clear. My apologies for that.</p> <p>If you want to see the demo please just drop me a line and i'll do the rest. At some point we intend to provide an unguided demo (maybe in the form of a series of <a href="http://www.qarbon.com/products/viewlet/">Viewlets</a>).</p> http://matt.blogs.it/entries/00001830.html So Dean lost Iowa Tue, 20 Jan 2004 06:51:22 +0000 <p>I saw the headline when I went to <a href="http://w4.evectors.it/">K-Collector</a> this morning and then read most of the details at the <a href="http://w4.evectors.it/itentdirectory/topic?topic=howard_dean">Howard Dean</a> topic. It made a very nice summary of the night and the various viewpoints.</p> http://matt.blogs.it/entries/00001842.html Permission based posting Sun, 08 Feb 2004 00:13:18 +0000 <p>Just a couple of follow-on thoughts...</p> <p><ol><li>If a feed is copyright, does that mean an aggregator should not collect it? Is this <em>reproducing</em> the work? (How is it different, for example, from downloading pirated software?)</li> <li>What does it say when an author asserts copyright of material, but publishes it in an RSS feed? The Creative Commons licenses grant rights to the recipient. Copyright (AFAIK) does not.</li> <li>Does it matter what type of aggregator it is? For example <a href="http://www.evectors.com/itkcollector/">K-Collector</a> aggregates in a different way (and for a different purpose) to, for example, <a href="http://radio.userland.com/">Radio Userland</a> or <a href="http://ranchero.com/netnewswire/">NetNewsWire</a>.</li> <li>If it aggregates a feed, should it prevent you from re-posting aggregated content? In Radio this is very easy, I can click a button on a post in the aggregator to use it as the basis for a post to my own weblog. If an item is copyright should it stop me doing that?</li> </ol></p> <p>I think this draws us awfully close to the kinds of things Ted Nelson was thinking about when he outlined his vision for transclusion based publishing. I can almost envisage a system where, when I press the post button on a copyright item, my aggregator goes off to check with the original authors system for permission to publish. You could imagine using the data from a social network to decide who can republish your content with your permission.</p> http://matt.blogs.it/entries/00001859.html Licensed to aggregate (pt #2) Tue, 10 Feb 2004 08:00:08 +0000 <p>I have received a couple of comments to my earlier <a href="http://matt.blogs.it/2004/02/08.html#a1315">piece</a> about licensing & RSS. Among others Phil Ringnalda pointed me at the <a href="http://backend.userland.com/creativeCommonsRssModule">Creative Commons RSS Module</a> authored by <a href="http://www.scripting.com/">Dave Winer</a> in Dec 2002.</p> <p>The CC module is pretty much the same solution as I was proposing in that it seems to be a start in the right direction. However neither it, nor my own suggestion, answer all the questions in a way I can appreciate (e.g. as Phil points out, what is covered by the license: text only? images? etc...), so i'm going to keep chipping at this.</p> <p>A commentor, <a href="http://www.lulop.com/">Lorenzo</a>, has suggested that <blockquote>Copyright is a "state" and licensing is an "action" made possible by the state.</blockquote>I kind of disagree with this statement but mainly on terminological grounds. I hope we can cut through that by agreeing that the central point is <em>rights</em>. Who has them? What uses do they permit?</p> <p>It seems to me that a Copyright statement reserves all rights relating to making copies (& derivative works) to the author who can then make exceptions on a case-by-case basis. As in the case where an author grants their publisher the right to make copies for sale.</p> <p>Hence an RSS feed with a copyright notice shouldn't be read (copied) unless you consider the act of offering feed to be an implicit agreement by the author to do so. Of course from a software perspective implicit rights can be problematic, especially when they are not immutable or well understood. For an RSS feed with a copyright statement what rights are actually on offer? People commonly republish content from posts they have aggregated either whole or in part. How do they know they have the right to do so?</p> <p>Then you have some interesting anomalies when software gets in the way. For example, my <a href="http://matt.blogs.it/">weblog</a> has a <a href="http://creativecommons.org/">Creative Commons</a> <a href="http://creativecommons.org/licenses/by-nd-nc/1.0">license</a>, yet my <a href="http://matt.blogs.it/rss.xml">feed</a> is Copyright. I guess Radio is automatically adding the copyright notice, I don't know how to make it stop. What am I telling you?</p> <p>Alternatively, and I guess this may be common, Phil Ringnalda's blog doesn't have any kind of license at all, neither does his feed. Can I assume Phil intends all his material to be public domain? If i'm not clear that this is the intension how can my software be?</p> <p>Creative Commons offer a range of licenses which offer the right to copy, or make derivative works, with certain restrictions such as <em>share alike</em> and <em>non commercial use only</em>. Primarily Public Domain takes this a step further in granting unlimited rights with exceptions being, if you'll excuse the pun, the exception.</p> <p>Maybe it's my developer "tunnel vision" at work but this looks very similar to the common model of permissions adopted in software everywhere: <ul> <li>include * except A, B, C,...</li> <li>exclude * except R, S, T,...</li> </ul> where A, B, etc.. are not users or hosts but <em>specific actions, by identified individuals, in specific circumstances.</em></p> <p>For most personal news aggregators I guess none of this matters much. If someone publishes a feed, and you're just reading it, then whats the harm? Unless of course you weren't supposed to have the URL to the feed. But that's a different issue.</p> <p>However for aggregators like <a href="http://w4.evectors.it/">K-Collector</a> and users who are reposting content it's a different matter. For example K-Collector doesn't mess with the content of posts, but it does republish them in a new context. If K-Collector has a better Google page rank than the author then we even begin to suck traffic away from them based upon their own material!</p> <p><em>(To be clear I am only talking about the public K-Collector portal <a href="http://w4.evectors.it/">W4</a>. This is an untypical use of <a href="http://www.evectors.com/itkcollector/">K-Collector</a> which is designed for use within organisations. But the point still stands. And what about <a href="http://www.feedster.com/">Feedster</a>?)</em></p> <p>As ambitious, semi-automated, software like K-Collector become more common then a reasonable, dependable, system of rights is going to be required. To my way of thinking the <a href="http://backend.userland.com/creativeCommonsRssModule">Creative Commons RSS Module</a>, whilst a start in the right direction, addresses a necessary, but not sufficient, subset of the goals. What do we do to take it further?</p> http://matt.blogs.it/entries/00001866.html ESF, topics and K-Collector Wed, 25 Feb 2004 22:08:05 +0000 <blockquote>Event Sharing, publishing, syndicating, etc. When we introduced the "when" part in the <a href="http://w4.evectors.it/">w4</a> concept, almost one year ago, what we had in mind was a space where events would be topics which could be aggregated in calendars and allowed users to navigate information using a timeline.<br> ... <br> Using the same approach we are using with topics, new events will be automatically distributed among members of a cloud allowing users to pick an event if it already exist instead of creating it. <br> <br> There will also be relations between events and other topics on the server, which we believe will create a sigificant added value to the process (allowing, for example, to quickly move to all information related to an event to all information related to one of the participants).<br> <br>[<a href="http://paolo.evectors.it/">Paolo Valdemarin: Paolo's Weblog</a>]</blockquote> <p>The <em>When</em> classification was always a key part of the overall vision of K-Collector. Being able to navigate sensibly based upon time & date is a very powerful concept.</p> <p>One of the last things I did within liveTopics before moving on to K-Collector was to implement time as an <a href="http://www.xfml.org">XFML</a> facet. This created a hierarchy of topics representing dates. For example there was a topic "2003". This in turn contained 12 sub-topics "Jan 2003" through "Dec 2003". Each of these topics contained a further division by date. Posts were linked to these topics based upon publication date. The upshot was that you could browse based on an increasingly specific date filter.</p> <p>We will be looking to implement <em>When</em> topics in K-Collector in conjunction with ESF events. The idea will be to link the event (perhaps defined as a <em>What</em> topic), with a particular point in time (a <em>When</em> topic), a place (<em>Where</em>) and possibly people attending (<em>Who</em>). Hence an event will form a glue which binds together a number of different topics in a context.</p> <p>Example: <dl> <dt>What</dt> <dd>blogtalk 2.0</dd> <dt>Where</dt> <dd>Vienna</dd> <dt>Who</dt> <dd>Thomas Burg</dd> <dt>When</dt> <dd>5-Jul-2004</dd> </dl></p> <p>There are still issues to work out like: What about events which span multiple days? Do we represent time on the calendar? What about recurring events? And so on. But I think even a simple model which dodges many of these questions would be amply useful at this point.</p> http://matt.blogs.it/entries/00001924.html More on #kmtalk and IRC integration Tue, 13 Apr 2004 17:58:47 +0100 A couple more visitors to <a href="http://matt.blogs.it/2004/04/09.html#a1399">#kmtalk</a>.&nbsp; First there was the author of <a href="http://jcwinnie.us/MT/weblog/">Your Guess Is As Good as Mine</a> (Jonathan Smith?).&nbsp; Also <a href="http://www.terryfrazier.com/weblog/">Terry Frazier</a> and <a href="http://www.rklau.com/tins/">Rick Klau</a> stopped by.&nbsp; Terry demonstrated his l33t skills by signing in using his Treo 600 (using <a href="http://www.smittyware.com/palm/upirc/">upIRC</a>).<br> <br> Anyone and everyone who is interested in knowledge management, communities of practice, collaboration, wiki, social networks and other related topics is welcome to stop by.&nbsp; Right now I am interested in how IRC can be linked to other tools (including <a href="http://www.evectors.com/itkcollector/">K-Collector</a> of course).<br> <br> For example might it be useful to link K-Collector topics (e.g. <a href="http://w4.evectors.it/itentdirectory/topic?topic=knowledge_management&chunck=1">Knowledge Management</a>) to IRC channels?&nbsp; For example topics like <a href="http://w4.evectors.it/itentdirectory/topic?topic=knowledge_management&chunck=1">Knowledge Management</a>, <a href="http://w4.evectors.it/itentdirectory/topic?topic=k_logs&chunck=1">K-Logs</a>, <a href="http://w4.evectors.it/itentdirectory/topic?topic=knowledge_organisation&chunck=1">Knowledge Organisation</a>, <a href="http://w4.evectors.it/itentdirectory/topic?topic=knowledge_work&chunck=1">Knowledge Economy</a>, <a href="http://w4.evectors.it/itentdirectory/topic?topic=collaboration&chunck=1">Collaboration</a>, <a href="http://w4.evectors.it/itentdirectory/topic?topic=commercial_blogging&chunck=1">Commercial blogging</a>, ... could all be linked to the #kmtalk IRC channel. Topics about different subjects could be linked to other channels.<br> <br> The idea is that people viewing the web topic could see who was talking in the channel (want to join in?) and recent traffic in the channel (have something to add?) Could this be a useful application?<br> http://matt.blogs.it/entries/00002073.html Getting Real with Stowe Tue, 20 Apr 2004 21:46:17 +0100 Had a first chat with <a href="http://www.corante.com/getreal/">Stowe Boyd</a> this evening.&nbsp; Enjoyed it very much despite some troubles with Skype (poor voice quality in one direction or another at various points) which we augmented with IM.&nbsp; We talked about <a href="http://www.evectors.com/itkcollector/">K-Collector</a>, IRC, Wiki, Sense making, Cynefin, social tools and how KM is suffering from <span style="font-style: italic;">fatigue</span>.&nbsp; Thanks Stowe!<br> http://matt.blogs.it/entries/00002088.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 <b>People Centred Knowledge Management</b> (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.<br> <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)<br> </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 <b>what I said</b>).</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. <br> <br> 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?<br> http://matt.blogs.it/entries/00002138.html Teenage Taxonomies Sat, 01 May 2004 10:40:23 +0100 I've been thinking about how topics are organized in <a href="http://www.evectors.com/itkcollector/">K-Collector</a>. As of my writing the largest K-Collector site is <a href="http://w4.evectors.it/">W4</a> which currently has 726 topics defined in 3 classifications.<br> <ul> <li>What - 410 topics<br> </li> <li>Who - 227 topics<br> </li> <li>Where - 89 topics<br> </li> </ul> This is beginning to present some challenges. For example current interfaces in K-Collector clients for selecting topics highlite suggested topics but also display all available topics (under their classifications) for you to choose from. The problem is that, with this number of topics, it's all rather unwieldy.<br> <br> I already think that it would be advantageous to allow a further subdivision of the classifications to form a 3-level hierarchy. For example What could subdivide into things like products, protoocols, principles, patterns and so on (sort of modelled after Denham Grays <a href="http://denham.typepad.com/km/2003/11/information_gat.html">Information Gathering Template</a>). But this is not a complete answer.<br> <br> First I do not want to grow arbitrarily nested taxonomies. If you think about your own experience with menu bars, how often do you look more than 3 levels deep? And how irritating is it to have to poke around like that? Also the deeper the taxonomy the more effort has to be put into designing it and this is the domain of experts and to be avoided<br> <br> Second faceted classifications only work when they narrow things. As you descend the hierarchy you become more precise about the term you are talking about (Anything => any person => a member of a group => an individual). But for example, when I am writing a post about politics and economics I am really only interested in topics related to those subjects, i.e. I have cross-cutting concerns (like <a href="http://server2.hostvalu.com/pipermail/discuss_aosd.net/2002-January/000052.html">Aspects w.r.t. OOP</a>) <br> <br> I've been thinking a little bit about Peter Van Dijck's <a href="http://poorbuthappy.com/ease/archives/002822.html">suggestion</a> for <i>nested facets</i>. As Travis Wilson described it in a post to the facetedclassification list:<br> <blockquote> <pre wrap="">For example, Peter's article poses a "Product Type" facet where "Cameras" <br>is a heading down in the taxonomy somewhere. Cameras have certain <br>properties like "Resolution" and "Lens Type" that just aren't relevant to, <br>say, hubcaps. So "Resolution" and "Lens Type" are facets with a scope of <br>"Product Type = Camera". A faceted navigation interface would expose them <br>whenever a search was already restricted to "Camera". Otherwise, they're <br>structured like every other facet you've ever seen.</pre> </blockquote> to see whether there is a role for nested facets. I'm still thinking about that.<br> <br> It's also possible that we could also use the many relations (thousands and thousands of them) that K-Collector builds up to create dynamic cross-cutting hierarchies. The idea here would be to take one or two dominant topics and then order all other topics according to how relevant they were. This should, in theory, put more relevant topics closer to your attention.<br> <br> No solutions yet, just questions & ideas.<br> http://matt.blogs.it/entries/00002158.html What ever happened to the Event Share Framework? Sat, 01 May 2004 11:13:54 +0100 The <a href="http://www.esfstandard.org/">Event Share Framework</a> seemed to have some buzz a little while back.&nbsp; I posted a question about implementation (we really wanted to do something like this in <a href="http://www.evectors.com/itkcollector/">K-Collector</a>) to the ESF list but got zero response and can't say I've seen anyone talking about it since.&nbsp; Did it disappear off the radar?&nbsp; Has it gone underground?&nbsp; Does anyone know?&nbsp; Or care?<br> http://matt.blogs.it/entries/00002168.html 800 topics bouncing around Wed, 12 May 2004 09:03:13 +0100 <blockquote><a href="http://www.community-intelligence.com/blogs/public/archives/000264.html">Visualizing social networks</a>. Increasing the capacity of people and communities to visualize their (online) social networks is essential to the evolution of Collective Intelligence. [<a href="http://www.community-intelligence.com/blogs/public/">Blog of Collective Intelligence</a>]<br> </blockquote> <br> I still really want some kind of visualization for K-Collector.&nbsp; We're generating lots of data, relationships between topics and posts, topics and topics, people and posts, people and topics, people and people and more all the time.&nbsp; When all you have is tables full of data it's hard to work out what might be interesting or valuable.<br> <br> The <a href="http://prefuse.sourceforge.net/">prefuse visualization library</a> looks very neat although I have done similar things in the past with <a href="http://www.touchgraph.com/">Touchgraph</a>.&nbsp; However a key problem remains, when dealing with large data sets, how to make a meaningful subset of them usable.&nbsp; Do you really want a graph with 800 topics bouncing around?<br> http://matt.blogs.it/entries/00002230.html K-Collector client demystified Thu, 20 May 2004 08:49:02 +0100 <blockquote><a href="http://www.readwriteweb.com/archives/001877.html">More on Topic-Sharing Community</a>. <p>There's already been a great response to <a href="http://www.readwriteweb.com/archives/001876.html">my post last night</a> (see the comments to previous entry). <a href="http://www.blogdigger.com/blog/index.html">Greg</a> suggested his aggregator <a href="http://www.blogdigger.com/">Blogdigger</a> could be included in this - I agree! <a href="http://matt.blogs.it/">Matt</a> and <a href="http://www.andrewsw.com/news/index.php">Andrew</a> also posted very thoughtful responses.</p> <p>Here's some of my feedback (copied from the comments - I must get these enabled inline...):</p> <p>Overnight while pondering my post (which I regard as just a 'starter for 10' btw, not a final solution by any means), I did conclude that KC essentially already does what I describe - polls registered RSS feeds with ENT in them and aggregates them. It would be great if TE also had that functionality.</p> <p>It's the client ping that I think is unnecessary and possibly holding back community uptake - with TE the ping is a manual process for the blogger, and with KC you need to install an add-on tool to enable the pinging. Both require too much manual effort for the blogger (IMHO of course). eg Bloglines does all its aggregation automatically (every hour I think), with no pinging required from the blogger.</p> <p>Although Andrew I take your point about bandwidth utilization. But if Bloglines (and Blogdigger) can do it, why not KC and TE?</p> [<a href="http://www.readwriteweb.com/">Read/Write Web</a>]<br> </blockquote>There seems to be some confusion about how <a href="http://www.evectors.it/itkcollector/">K-Collector</a> server works and the role of the K-Collector client so I thought I would try and give an explanation of how things fit together.<br> <br> The first thing to understand is that you absolutely <b>do not</b> need the K-Collector client for your blog to be part of a K-Collector site. The client offers a set of benefits aimed at improving the experience for the user, but they are entirely optional. We currently aggregate many feeds to the <a href="http://w4.evectors.it/">W4</a> site which are not using one of our clients.<br> <br>There are three reasons why we think using the client is beneficial:<br> <br> 1) The ping (it's the least important, but seems most misunderstood so I'll cover it first)<br> <br> The K-Collector server contains an aggregator which reads all feeds on a rotating basis. It aims to read each feed more or less once per hour but this isn't guaranteed. It collects posts from feeds and assigns them to topics using either <a href="http://www.purl.org/NET/ENT/1.0/">ENT</a> metadata supplied in the feed or choices which are auto-discovered using various word-stemming and matching techniques.<br> <br> This all happens entirely independent of the client ping. <b>All the ping does is to move your feed up the list so that new posts you have written are likely to be collected sooner.</b> If you don't ping the server just reads your feed automatically a little later on, that's all.<br> <br> 2) The topic manager<br> <br> Through the client, the topic manager is integrated into the blog editing process and gives authors the ability to assign community topics to their posts as well as being able to create new topics. The topic manager also attempts to suggest topics which may be relevant to the content of the post to make choosing topics easier.<br> <br> Without the client you have no way to decide which topics should be assigned to each post. In this case the server will, when it reads the feed, use it's own automatcher to automatically assign those topics it thinks are relevant.<br> <br> 3) ENT feeds<br> <br> Where the author has choosen topics in the topic manager the client adds the appropriate ENT metadata to the outgoing RSS feed. K-Collector can then use this metadata to accurately assign posts to topics.<br> <br> In summary the K-Collector client offers what we think are very useful benefits to weblog authors, however it is entirely optional and you do not need it for you weblog to be part of a K-Collector site. Equivalently K-Collector itself only cares about RSS.&nbsp; It doesn't care whether ENT metadata was created by our client or some other application, we're completely agnostic about that.<br> <br> I hope this goes some way to clearing up how the K-Collector system works.<br> http://matt.blogs.it/entries/00002257.html Getting my head shifted Tue, 25 May 2004 09:29:09 +0100 Had a very nice lunch with Lee Bryant of <a href="http://www.headshift.com/">HeadShift</a> yesterday.&nbsp; We met in <a href="http://matt.blogs.it/2004/04/09.html">#kmtalk</a> talking about the <a href="http://www.socialtext.net/stes/">Social Tools for Enterprises Symposium</a> I am helping to create..&nbsp; Headshift are based in <a href="http://www.pooloflondon.co.uk/visiting.builder/places/0010.html">Butlers wharf</a>, right on the river, which is a great setting for a meet up.<br> <br> Over spinach & ricotta parcels and some <a href="http://www.staropramen.cz/">Staropramen</a> we chatted about: social software, the challenges facing organisation and employees, the central importantance of people at all stages of collaborative/KM projects, the roles of blognets, <a href="http://www.headshift.com/work/proj.cfm">projects</a> they are working on, <a href="http://www.foaf-project.org/">FOAF</a> (esp. how we need to be careful with what we infer from the data), <a href="http://www.sysval.org/">metavalues</a>, <a href="http://www.linkedin.com/">LinkedIn</a>, <a href="http://matt.blogs.it/2003/06/19.html#a956">Dave Snowden</a>, <a href="http://www.evectors.com/itkcollector/">K-Collector</a>, <a href="http://www.purl.org/NET/ENT/1.0/">ENT 1.0</a>, <a href="http://blogtalk.net/">BlogTalk 2.0</a>, the <a href="http://topicexchange.com/t/blogwalk/">BlogWalkers</a>, Vienna, services vs. products, and a host of other topics.<br> <br> I look forward to talking to Lee & co. again in the near future.<br> http://matt.blogs.it/entries/00002276.html Tracking ENT support Wed, 09 Jun 2004 11:21:27 +0100 I've just made a minor change to the <a href="http://www.purl.org/NET/ENT/1.0/">ENT 1.0 spec</a> to add an <a href="http://www.purl.org/NET/ENT/1.0/index.html#implementations">implementations section</a>. Obviously I've listed <a href="http://www.evectors.com/itkcollector/">K-Collector</a> as the first implementation would like to gather together references to anyone who has supported the format in some way. So, if you know of any software or service which is supporting, or working on supporting, ENT please let me know so that I can add it to the spec document.<br> <br> Update: <a href="http://www.myelin.co.nz/post/">Phil Peasons</a> <a href="http://topicexchange.com/">Internet Topic Exchange</a>, of course!<br> http://matt.blogs.it/entries/00002308.html Mon, 05 Jul 2004 22:15:06 +0100 In <a href="http://www.blogtalk.net/">BlogTalk</a> 2.0 <a href="http://randgaenge.net/">Thomas</a> & crew have, once again, brought together a lot of interesting people & wherever possible conversations are flourishing as evidenced by tonights get together. It was good to meet <a href="http://www.dijest.com/aka/">Phil</a> again after a year and to meet <a href="http://radio.weblogs.com/0100875/">Mikel</a> for the first time, especially as his talk prompted lots of interesting ideas we might look at for <a href="http://www.evectors.com/itkcollector/">K-Collector</a>. After a day where I had felt very tired and jaded I found the atmosphere quite reviving.<br> <br> I had the pleasure of dining with <a href="http://paolo.evectors.it/">Paolo</a>, <a href="http://www.markbernstein.org/">Mark Bernstein</a>, <a href="http://www.sumofmyparts.com/blog/">Stephanie Hendrick</a>, and <a href="http://blog.humlab.umu.se/therese">Therese Örnberg</a>. Mark gave a very interesting keynote this morning which provoked lots of questions for me. Stephanie & Therese gave, I think, the most stylish presentation of the day (including an amusing near-death audioblog to end) and their discussion of presence and spaces was stimulating. From my perspective a happy coincidence that we all ended up together. We had an interesting discussion about a range of topics spanning language, blogging, literary discourse, topics, flame wars, comments & trackbacks, software tools and how you build them, tinderbox, Dave Allen, and test first development.<br> <br> Taking antibiotics means I cannot drink alchohol so my opinions whilst maybe better formed were far less robust than usual & I was open to colonization ;-) I got persuaded that comments are bad and that even trackback requires considerable architectural revision to work properly. Mark's suggestion of making trackback default to being private (i.e. you get a file of trackbacks and you decide what, if anything, to do with them) seems to be a good one. I think this can be assisted by some sort of intelligent filtering of trackback contents & authorship to help you decide about those you do & don't want to handle. I think emulating the LinkedIn FOAFOAFOAFOAF network model could be useful in this regard.<br> <br> Also, based on comments by Stephanie and Mark, I have finally concluded that I must do something in K-Collector for the <a href="http://blog.mathemagenic.com/">Lilia's</a> and <a href="http://www.mcgeesmusings.net/">Jim McGee's</a> of this world who used (and maybe still cling to) <a href="http://www.google.com/search?num=20&hl=en&lr=lang_en&ie=UTF-8&safe=off&c2coff=1&q=livetopics&btnG=Search">liveTopics</a>. I think that part of my problem has been misunderstanding where they are coming from. liveTopics, for me, was a stepping stone towards a larger vision which, at that time, I couldn't achieve. But for them it was actually what they were looking for. No wonder then that I've had a hard time convincing them that K-Collector is better.<br> <br> I haven't quite worked out the answer yet but I think it may be as simple as offering some kind of discriminator where you can choose whether K-Collector should default to showing you only your own work, or the work of the community as a whole. We may even have enough smarts in the database to do this without requiring additional work but I'll have to get some clear space (i.e. after <a href="http://stes.evectors.com/">STES</a>) to think this through properly.<br> http://matt.blogs.it/entries/00002390.html Visualizing concepts in weblogs Tue, 06 Jul 2004 13:10:30 +0100 <a href="http://anjo.blogs.com/">Anjo Anjewierden</a> & <a href="http://blog.mathemagenic.com/">Lilia</a> are demonstrating a very interesting tool for visualizing conceptual relationships within weblogs.&nbsp; It uses technology similar, but more advanced, to K-Collector for analyzing the content of weblog posts and drawing out concepts which can be analyzed and displayed as a network so that you can explore the relationships.&nbsp; It also allows comparison of terms between different weblogs.<br> <br> Very intriguing!<br> http://matt.blogs.it/entries/00002404.html DelphiGroup: Making the case for taxonomy Mon, 16 Aug 2004 11:56:57 +0100 I'm reading an excellent report from <a href="http://www.delphigroup.com/">Delphi Group</a> called Information Intelligence: <i>Content Classification and the Enterprise Taxonomy Practice</i> (see <a href="http://www.delphigroup.com/coverage/taxonomy.htm">here</a> for a download link). The gist of the report is that, for enterprises, search is not enough.<br> <br> The report makes the case that whilst search technology has improved a lot in recent years, and continues to improve, the majority of professionals still find it an unsatisfactory way to work and often spend in the region of 20% of their time searching for information. Often cited problems were constantly changing information and a lack of precision about what they were looking for. The report then makes a case that introducing taxonomy based services can significantly improve performance and save money by eating into that 20%.<br> <br> Whilst the report is funded by a number of companies with a vested interest in taxonomy (for example Autonomy or Verity) the case seems to be well made off the back of a credible piece of research (which is a follow-up to similar research done last year).<br> <br> My summary:<br> <ul> <li>82% of users do not have access to a centralized point of search & information across information systems</li> <li>The provision of a singular navigational front end (e.g. taxonomy) and omnipresent search tool that collectively aggregate disparate content resources, can, from an end-user perspective, deliver the simple single point of access that many users strive for.</li> <li>Lack of organisation of information is the number one problem in information management & retrieval.</li> <li>If professionals are spending 20% of their time (or more) looking for information then this results in an opportunity cost & represents a runaway expense item in many organisations.</li> <li>Keyword search assumes you know what you are looking for & that it an often erroneous assumption.</li> <li>75% of people surveyed during a Yahoo market research project preferred browsing to searching.</li> <li>In some instances it is easier to discover information about a particular subject if you see it in the context of related thought. Browsing encourages associative thought.</li> <li>The availablity of taxonomy eliminates the need for the researcher to completely understand the subject before issuing a query.</li> <li>Browsing via a taxonomy in essence provides an education on the subject and lends insight into the issues or facets of the subject.</li> <li>The number one source of frustration with search of on-line content is the fact that the content they search for is constantly changing, which both frustrates the user and reduces the effectiveness of simple search.</li> <li>Use of a taxonomy can provide a dynamic bookmark so to speak, a one-stop-shopping guide to all relevant content on a subject. Return to a subject node exposes the latest and complete collection of content about that subject area. This addresses the number one cause of frustration, the dynamic, volatile nature of information sources.</li> <li>Navigation of a well-designed interface to information on a web site/portal automatically directs the researcher to other relevant topics.</li> <li>The tagging effort represents another process that a business must undertake in order to obtain the benefits of a taxonomy. In some cases this could be done manually. But this approach is not easily scalable.</li> <li>Will authors be willing or available to perform this classification manually?</li><li>68% concur that the process of locating & retrieving the information needed to effectively execute their jobs is difficult and time consuming.&nbsp; Not a single respondent strongly disagreed with this statement.</li> <li>Whilst users see some improvement in information retrieval over the last 2 years, their attitude towards its level of difficulty remained virtually the same.</li> <li>Respondents overwhelmingly pointed to the fact that business content is constantly changing and has to be continually relocated.</li> <li>Findings pointed more towards the speed and ease of use of retrieval environments and less to effectiveness, as the primary point of pain amongst todays business people.<br> </li> </ul> I think this report is very well worth reading to anyone interested in search, taxonomy, or knowledge organisation. Of course I too am baised because I think <a href="http://www.evectors.com/itkcollector/">K-Collector's</a> integrated approach addresses a number of the concerns raised by this report.<br> http://matt.blogs.it/entries/00002446.html