Curiouser and Curiouser!

Curiouser and Curiouser! on ent-1.0 http://matt.blogs.it/ RSS feed for topic ent-1.0 Copyright 2006 Matt Mower Squib/0.1 self@mattmower.com self@mattmower.com en-gb Wanna know how it works? Tue, 13 May 2003 20:38:08 +0100 <BLOCKQUOTE dir=ltr style="MARGIN-RIGHT: 0px"> <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>]</BLOCKQUOTE> Yeah!  That's how it works. http://matt.blogs.it/entries/00001464.html When aggregators attack Mon, 18 Aug 2003 14:28:23 +0100 <BLOCKQUOTE dir=ltr style="MARGIN-RIGHT: 0px"> <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>]</BLOCKQUOTE> The title of the article  Aggregators Attack Info Overload  suggested to me a slightly more in depth piece looking at the value of using RSS and aggregators to communicate and share.  In this, I think it fell short. I'm also disappointed because it meant that <A href="http://w4.evectors.it/">K-Collector</A> didn't get a mention.  We may be the new kid on the block but with RSS+ENT I think we're doing something really interesting in this space. http://matt.blogs.it/entries/00001551.html New German language blogging directory Sun, 24 Aug 2003 16:26:31 +0100 <A href="http://www.lumma.de/">Nico Lumma</A> just pinged me to tell me about his <A href="http://www.blogworkorange.de/">companies</A> new blog listing site <A href="http://www.blogg.de/">BLOGG.de</A>.  The site is building a categorized directory of German language blogs and they are using a combination of <A href="http://www.movabletype.org/trackback/">trackback</A> and <A href="http://matt.blogs.it/specs/ENT/1.0/">ENT 1.0</A> to do it. Very cool. http://matt.blogs.it/entries/00001577.html Considering RSS-Data Wed, 08 Oct 2003 23:09:56 +0100 I was going to leave this as a comment on <a href="http://www.docuverse.com/blog/donpark/2003/10/08.html#a961">Don's piece on RSS-Data</a> (hello Don!) but it grew to the point where a post felt more appropriate.  Anyway my grateful thanks to Don, for his short but sweet explanation. <div style="margin-left: 40px;">RSS-Data = Typed maps for RSS2.0 right? </div> Okay i'm glad it's an RSS2.0 extension.  But is it a good thing? I guess I have bought into the idea that one of the things we bought when we paid for XML was the notion that the tags carry semantic value. Applications having to understand the semantics is one of the prices of doing something useful.  Or at least that was what I thought. But maybe that shouldn't be the case for things such as extensions.  I guess i'm wondering whether <a href="http://matt.blogs.it/specs/ENT/1.0/">ENT</a> would have been written as a separate module if RSS-Data had been around in April.  Maybe it would have been easier to simply identify a few RSS-Data entries to add to an item to convey topic information...? My gut instinct is that it might have been easier to sell, but that it wouldn't make as much sense.  Describing ENT tags strictly as a set of attribute:value pairs would be more confusing and possibly harder to process.  I think ENT has a distinct flavour and that this is not a bad thing. There's room for both approaches, we'll just have to see. Oh and a caveat:  Your data isn't going to be named signature, value and cert unless you want them clashing with any other possible applications of those names, so you'll probably end up prefixing them as in: <ul> <li>com.docuverse.blah.signature</li> <li>com.docuverse.blah.value</li> <li>com.docuverse.blah.cert</li> </ul> http://matt.blogs.it/entries/00001663.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>] </div> We're publishing several hundred topical 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>.  You can navigate the feed structure quite easily.  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.  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. http://matt.blogs.it/entries/00001672.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>] </div> <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.  That's great! 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.  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. But why stop there.  Your junit based daily smoke & build test could be generating a similar feed, your project management system, issue database and so on.  All these feeds could be flowing into project topics giving you an uptodate and holistic view of what is happening in those projects. Who'd ever be caught on the hop in a project meeting again? http://matt.blogs.it/entries/00001747.html Enhancing Radio's aggregator Wed, 25 Feb 2004 21:49:25 +0000 <blockquote><a href="http://radio.weblogs.com/0104487/outlines/radio/enhancedAggregator.html">Introducing the enhancedAggregator</a>. <a href="http://radio.weblogs.com/0104487/images/my/aggregator.png" target="_blank" title="view hires"><img src="http://radio.weblogs.com/0104487/images/my/aggregatorLR.gif" alt="aggregator topics" align="right" border="0"></a>I've recently spent some time investigating Radio's aggregator code, looking for an easy way to support additional RSS modules in general, and ENT 1.0 topics in particular. The <a href="http://updates.prec-it.com/download/enhancedAggregator.root" title="click to download">enhancedAggregator tool</a> is the -provisional- result of this investigation. It comes with full <a href="http://matt.blogs.it/specs/ENT/1.0/" target="_blank">ENT 1.0 topics</a> support for Radio's aggregator, and skeletons for aggregating <a href="http://www.intertwingly.net/wiki/pie/FrontPage" target="_blank">Atom 0.3 feeds</a> and <a href="http://www.esfstandard.org/" target="_blank">ESF 1.0 events</a> for RSS 2.0. I'd like the enhancedAggregator to become a community driven project, allowing Frontier/Radio developers to easily test aggregation of new syndication formats and extensions, without mobilizing Userland scarse resources. So I've added some intal/uninstal/update/prefs ancillary functions to the tool, and provided <a href="http://radio.weblogs.com/0104487/outlines/radio/enhancedAggregator.html" target="_blank">guidelines</a> for updating the current drivers and adding new ones, with pointers to the available online documentation. I hope <a href="http://matt.blogs.it/" target="_blank">Matt</a> will copy the <a href="http://matt.blogs.it/specs/ENT/1.0/" target="_blank">ENT</a> module driver and paste it into the <a href="http://groups.yahoo.com/group/k-collector-support/" target="_blank">k-collector</a> client for Radio, and <a href="http://paolo.evectors.it/2004/02/24.html#a2062" target="_blank">Paolo</a>'s eVector crew will build upon the <a href="http://www.esfstandard.org/" target="_blank">ESF</a> module driver skeleton, copying the result to their new tool when it's stable enough. [<a href="http://radio.weblogs.com/0104487/">s l a m</a>]</blockquote> Marc's doing sterling work. http://matt.blogs.it/entries/00001916.html We built a better mousetrap, where are the mice? Sat, 08 May 2004 22:53:24 +0100 <blockquote><a href="http://matt.blogs.it/specs/ENT/1.0/">Wouldn't it be cool?</a>. Wourldn't it be cool if otehr people started parsing the ENT tags embedded in this post? That way if I talked about FOAF - for instance - someone like danbri could scarf JUST the FOAF posts and do anything he wanted with them! This would create an incredible two-way kind of capability because systems could then communictae back to me based upon what I said. I know I know -it's RSS2.0 but that becomes a really nice gateway to a world that has 75% market share of feeds(maybe even more.) [<a href="http://blogs.it/0100198/">Marc's Voice</a>] </blockquote> I think it would be pretty cool -- I've been hoping other would start grokking this for a while now.  Yes it's RSS2.0, yes it's not perfect, but it's here, now and I think that by the 80/20 rule it's good enough.  If there's anything we can do to help get ENT support included in other applications please let us know and we will do our best to help. I'd also like to see applications start using the <a href="http://matt.blogs.it/specs/SGUID/1.0/">SGUID</a> information that Paolo and I have had in our feeds for about the same length of time as we've been doing ENT.   Using SGUID one post in an RSS feed can refer directly to the permalink of the post it is quoting from.  Standard RSS2.0 only allows you to refer to the feed.  So for example, tags like: <blockquote><sguid:sourceRef>http://www.theobviousblog.net/blog/archives/000506.html</sguid:sourceRef> </blockquote> should allow a clever aggregator to thread posts. An aggregator that did topics and sguid-based threading, that would be nice... http://matt.blogs.it/entries/00002204.html (ENT2.0 mod RSS1.0) = 0 Mon, 10 May 2004 10:51:09 +0100 <blockquote>Time for ENT 2.0?. It's very interesting to read Danny's toughts about <a href="http://dannyayers.com/archives/002586.html">ENT and RSS 1.0</a>. Maybe it's time for a new release of the ENT specs, RSS 1.0 compatible. Oh... and what about Atom? [<a href="http://paolo.evectors.it/">Paolo Valdemarin: Paolo's Weblog</a>] </blockquote> I've certainly thought about things which, with the benefit of hindsight, I would have done differently.  I was never comfortable with having the topic name as the text content of the <code><topic></code> element and I've no idea why I did it, there are other bugbears in there too. I'd also like to give more thought as to how ENT feeds can be supported by topic map resources in real applications.  At the moment we don't publish XTM or XFML maps out of K-Collector but we could (I used to publish XFML from liveTopics but those files got big!) Lastly I would really like to make a push for ENT support in other applications.  It seems a shame to me that, more than a year on, no other applications seem to have picked up on the benefits topic based aggregation offer to users. http://matt.blogs.it/entries/00002215.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>. 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. Here's some of my feedback (copied from the comments - I must get these enabled inline...): 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. 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. Although Andrew I take your point about bandwidth utilization. But if Bloglines (and Blogdigger) can do it, why not KC and TE? [<a href="http://www.readwriteweb.com/">Read/Write Web</a>] </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. The first thing to understand is that you absolutely do not 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. There are three reasons why we think using the client is beneficial: 1) The ping (it's the least important, but seems most misunderstood so I'll cover it first) 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. This all happens entirely independent of the client ping. 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. If you don't ping the server just reads your feed automatically a little later on, that's all. 2) The topic manager 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. 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. 3) ENT feeds 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. 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.  It doesn't care whether ENT metadata was created by our client or some other application, we're completely agnostic about that. I hope this goes some way to clearing up how the K-Collector system works. http://matt.blogs.it/entries/00002256.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.  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..  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. 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. I look forward to talking to Lee & co. again in the near future. http://matt.blogs.it/entries/00002273.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. Update: <a href="http://www.myelin.co.nz/post/">Phil Peasons</a> <a href="http://topicexchange.com/">Internet Topic Exchange</a>, of course! http://matt.blogs.it/entries/00002307.html All formats lead to Rome (and back again) Thu, 17 Jun 2004 22:04:17 +0100 I'm doing some work with <a href="https://rome.dev.java.net/">Rome</a> which is an ambitious open source Java RSS toolkit by <a href="http://blogs.sun.com/tucu/">Alejandro Abdelnur</a>, <a href="http://www.chanezon.com/pat/weblog/">Patrick Chanezon</a>, and Elaine Chien of Sun. Rome does not attempt to be all things to all men, choosing (correctly imo) to concentrate on providing low-level feed parsing & generation services.  It can read & write RSS in all it's flavours 0.9x, 1.0, 2.0 as well as Atom 0.3.  It can also convert feeds between formats and provide an abstract syndication feed layer over them.  Crucially they have good support for modular extension & I am bashing out an <a href="http://www.purl.org/NET/ENT/1.0/">ENT 1.0</a> module which I'll contribute back to the project. It's still alpha but already looks impressively useable.  Good stuff. http://matt.blogs.it/entries/00002338.html