« Brrr, yee hah, brrr, yee hah, brrrrr | Main | Inertia Injuries and Barr Blogging »
Matt 'BlackBelt' Jones has many good phrases, but his best is "LazyWeb": the idea that if you want to make something, just post about it and chances are people will do it first, or will have already done it. It works for him, and so I'll try it myself.
Here's the thing: I want to make a More Like This From Others button for each of the entries below. Clicking on it would bring a list of entries, formatted just like the blog, with excerpts of entries on a similar subject from other people.
Herein lies the rub: following my little investigation into people's categorisation habits, I can see that everyone calls everything different names: and until we have a universal taxonomy for weblogs (ha!) then we're stuck on a global "Google for Weblogs".
So here's what I'd like. Movable Type blogs now automatically create trackbacks when they can. These trackbacks contain RDF, denoting the category the MT blog has that category within. MT produces RDF indexes too (in the flavour of RSS 1.0). So, what I want is a little app that takes the trackback. Follows it back to the originating site, find the RDF snippet, takes the index.rdf, and gives back all the entries within the index.rdf that are on the same subject as the trackback one.
Got that?
Anyone got that already?
UPDATE I think this RDF Query Primer might help. Hmm.
This entry's trackback URL is http://blog.mediacooperative.com/mt-tb.cgi/1017
Ben Hammersley.com: Trackback, RDF, and the LazyWeb: So here's what I'd like. Movable Type blogs now automatically create trackbacks when
Posted at December 10, 2002 08:31 PM
Slashdot : Are Blogging and Unemployment Related? It would be interesting to see if there's a correlation between the meteoric...
"Unemployed blogger"
10 goldene Regeln des Webdesign nach: TU Chemnitz 10 goldene Regeln des Blogdesign nach: diversen Quellen und Meinungen Vermeide...
"Designregeln"
Thomas meint: Titel und Kategorien in Weblogs sind übrigens Kastration des Hausherrn, weshalb es hier nur eine Kategorie (täglich) und...
"Ohne Titel"
Ben Hammersley quotes Blackbelt Jones (a participant of the now infamous Shirkygate): Matt 'BlackBelt' Jones has many good phrases, but his best is "LazyWeb": the idea that if you want to make something, just post about it and chances are
Posted at December 11, 2002 10:51 PM
Ben Hammersley quotes Blackbelt Jones (a participant of the now infamous Shirkygate): Matt 'BlackBelt' Jones has many good phrases, but his best is "LazyWeb": the idea that if you want to make something, just post about it and chances are
Posted at December 11, 2002 10:52 PM
A possible solution to the problem of different names for categories: start from your TrackBack pings, get the category from the TB RDF, take that to the RSS feed, and get items in the same category. Twisty!
Posted at December 12, 2002 04:46 AM
Ben Hammerlsey wants more. Well "More Like this" to be precise: Movable Type blogs now automatically create trackbacks when they
Posted at December 13, 2002 10:00 AM
Ben Hammerlsey wants more. Well "More Like this" to be precise: Movable Type blogs now automatically create trackbacks when they
Posted at December 13, 2002 10:03 AM
Ben Hammerlsey wants more. Well "More Like this" to be precise: Movable Type blogs now automatically create trackbacks when they
Posted at December 13, 2002 11:50 AM
Ben Hammersley wrote a couple of days ago about Trackback, RDF, and the LazyWeb Here's the thing: I want to
Posted at December 13, 2002 12:56 PM
Article on how professional researchers can use blogs to advantage
"Personal Knowledge Publishing"
Prompted by Ben Hammersley's idea, Ben Trott has written the More Like This From Others script and Ben has implemented...
"Update on linking blogs by category"
Hammersley, Azhar and others grapple with linking blogs by category
"Categorising Blogs"
Link to initiative to provide a metadata standard for knowledge-logging
"Weblog metadata"
Experiences introducing K-logging into a company
"K-Logging Pilot"
Hammersley, Azhar and others grapple with linking blogs by category
Posted at December 13, 2002 12:58 PM
Article on how professional researchers can use blogs to advantage
"Personal Knowledge Publishing"
Prompted by Ben Hammersley's idea, Ben Trott has written the More Like This From Others script and Ben has implemented...
"Update on linking blogs by category"
Hammersley, Azhar and others grapple with linking blogs by category
"Categorising Blogs"
Link to initiative to provide a metadata standard for knowledge-logging
"Weblog metadata"
Experiences introducing K-logging into a company
"K-Logging Pilot"
Hammersley, Azhar and others grapple with linking blogs by category
Posted at December 13, 2002 01:00 PM
Article on how professional researchers can use blogs to advantage
"Personal Knowledge Publishing"
Prompted by Ben Hammersley's idea, Ben Trott has written the More Like This From Others script and Ben has implemented...
"Update on linking blogs by category"
Hammersley, Azhar and others grapple with linking blogs by category
"Categorising Blogs"
Link to initiative to provide a metadata standard for knowledge-logging
"Weblog metadata"
Experiences introducing K-logging into a company
"K-Logging Pilot"
Ben Hammersley's LazyWeb challenge to build a "More Like This From Others" is a fantastic use of TrackBack and RSS.
Posted at December 13, 2002 11:50 PM
Based on the endless stream of weblog press (and the limited number of unique angles), I've come to the conclusion...
"Press and Discovery"
Dan Gillmor, technology columnist for the Mercury News, previews his MT-powered weblog....
"Dan Gillmor's EJournal"
Ben Hammersley has posted an index template to support RSD for Movable Type's XML-RPC implementation. RSD (Really Simple Discoverability) allows...
"Support for RSD"
Ben Hammersley's LazyWeb challenge to build a "More Like This From Others" is a fantastic use of TrackBack and RSS....
"More Like This From Others"
Jeremy Zawodny states the need for "some sort of cross-weblog threading system" and questions whether this should be TrackBack. I've...
"TrackBack and Cross-Weblog threading"
pb posted a while back showing how PGP signatures could be integrated into comment systems. It's a slick idea. Of...
"Verifying PGP Signatures"
Geeks – at least those with any vision – dream of ‘the singularity’. For the social software geeks, this singularity
Posted at December 14, 2002 12:42 AM
Geeks – at least those with any vision – dream of ‘the singularity’. For the social software geeks, this singularity
Posted at December 14, 2002 12:46 AM
Geeks – at least those with any vision – dream of ‘the singularity’. For the social software geeks, this singularity
Posted at December 14, 2002 12:57 AM
Ben Trott has answered Ben Hammersley's "LazyWeb" quest. This is what Ben Hammersley wanted (and got): "So, what I want
Posted at December 14, 2002 07:44 AM
Ben Trott has answered Ben Hammersley's "LazyWeb" quest. This is what Ben Hammersley wanted (and got): "So, what I want
Posted at December 14, 2002 07:46 AM
Geeks – at least those with any vision – dream of ‘the singularity’. For the social software geeks, this singularity
Posted at December 15, 2002 12:22 AM
Prompted by Ben Hammersley's idea, Ben Trott has written the More Like This From Others script and Ben has implemented
Posted at December 15, 2002 01:43 PM
Article on how professional researchers can use blogs to advantage
"Personal Knowledge Publishing"
Prompted by Ben Hammersley's idea, Ben Trott has written the More Like This From Others script and Ben has implemented...
"Update on linking blogs by category"
Hammersley, Azhar and others grapple with linking blogs by category
"Categorising Blogs"
Link to initiative to provide a metadata standard for knowledge-logging
"Weblog metadata"
Experiences introducing K-logging into a company
"K-Logging Pilot"
Everytime someone in blogdom comes up with something new (FOAF, RSS link tag, blogchalking...) I wonder why I hadn't thought of it before, for most times they are great, simple ideas. Sometimes (like in CSS styling RSS feeds) I think to myself: well, I...
Posted at December 16, 2002 05:55 PM
Interesting article by Mark Pilgrim at XML.com: What is RSS?.
Now that I'm writing a tool that will consume RSS, I welcome this insight on how to deal with the myriad versions of RSS out there. For the moment I'm using XML::RSS, which means I'm restricted to RSS 1.0 feeds (I don't feel like supporting 2.0 for the moment, and the meat of the tool is displaying the time info on items, so 0.9x is not up to par, for the moment...)
Of interest also the commentary of the author on the process behind his appearance on XML.com
"RSS tips at XML.com"
I have been reading (and experimenting) a lot with SVG. It's an interesting technology, and has allowed me to discover many interesting sites, it seems SVG developers are very smart. I will try to use more and more SVG on this site,as I learn more; but within context, of course.
"SVG"
Everytime someone in blogdom comes up with something new (FOAF, RSS link tag, blogchalking...) I wonder why I hadn't thought of it before, for most times they are great, simple ideas. Sometimes (like in CSS styling RSS feeds) I think to myself: well, I had already thought of this, there was just some reason it didn't work for me and could'nt spend enough time experimenting with it to make it work.
OK, I have enough of this. I've been maturing an idea in the back of my mind, and have begun implementing it. But, because of this lazy web capriciousness, someone other could complete it before I get to it, so I'll just let this entry here as proof I had already thought of it:
I am writing something (probably an MT template, and if it can't be done, an MT plugin) to render a timeline of posts to a blog, in SVG. I'll use CPAN's SVG module and probably MTPerlScript to do it.
Of course, I need all my readers to blog this idea so as to certify I had already thought of it first ;)
"For the record: Weblog Timeline in SVG"
...is being forced to use it. And, even more than that, when using a strongly typed language such as C#, which has no need for such monstrosities.
I agree a strong adherence to coding conventions is essential, but who wants to use legacy coding conventions?
Lord, give me patience!!
"What I dislike more than Hungarian notation..."
Just when you think this blogosphere can't be any more connected, along comes yet another way to highlight, comment on
Posted at December 17, 2002 12:27 AM
I'm using SpamAssassin now on my machine to kill the crap that floats thorough from marketers. It seems to be...
"Strange Spam"
Just when you think this blogosphere can't be any more connected, along comes yet another way to highlight, comment on...
"More Connections"
Mandrake's moving into the post-install configuration, less than 20 minutes after starting the package install. Not bad. Adding users, including...
"Mandrake Install - III"
Package group selection is easy in Mandrake. The options for a workstation are Office Workstation, Game Station, Multimedia Station, Internet...
"Mandrake Install - II"
The lava lamp's fired up. The mp3 are playing randomly. And I've switched from espresso to Mello Yello. All's ready...
"Mandrake Install - I"
Mandrake's on the menu this afternoon. I've got a piping hot box of Mandrake Linux ProSuite 9.0 sitting on my...
"Mandrake Blogcast"
The Register is reporting that the Secret Service was used once again in a rather odd way. This time, it...
"Secret Service Again"
I'm living in a wireless cloud at home. My 2.4 Ghz Linksys access point and WCF11 CF card make the...
"WiFi Workers"
Here's an intesting use of Linux clustering. Saudi Aramco is building a 4,000 unit Linux cluster to look for oil....
"4,000 Systems"
Just when you think this blogosphere can't be any more connected, along comes yet another way to highlight, comment on
Posted at December 17, 2002 12:40 AM
I'm using SpamAssassin now on my machine to kill the crap that floats thorough from marketers. It seems to be...
"Strange Spam"
Just when you think this blogosphere can't be any more connected, along comes yet another way to highlight, comment on...
"More Connections"
Mandrake's moving into the post-install configuration, less than 20 minutes after starting the package install. Not bad. Adding users, including...
"Mandrake Install - III"
Package group selection is easy in Mandrake. The options for a workstation are Office Workstation, Game Station, Multimedia Station, Internet...
"Mandrake Install - II"
The lava lamp's fired up. The mp3 are playing randomly. And I've switched from espresso to Mello Yello. All's ready...
"Mandrake Install - I"
Mandrake's on the menu this afternoon. I've got a piping hot box of Mandrake Linux ProSuite 9.0 sitting on my...
"Mandrake Blogcast"
The Register is reporting that the Secret Service was used once again in a rather odd way. This time, it...
"Secret Service Again"
I'm living in a wireless cloud at home. My 2.4 Ghz Linksys access point and WCF11 CF card make the...
"WiFi Workers"
Here's an intesting use of Linux clustering. Saudi Aramco is building a 4,000 unit Linux cluster to look for oil....
"4,000 Systems"
I'm very intrigued by the lately discussions about More Like This From Others. Its such an interesting concept, but I'm
Posted at December 17, 2002 03:21 AM
The Web Standards Project comes back with a new look and some great new information. Do you have a standards...
"WaSP update"
I am so excited to see this - TopStyle Pro 3.0 Beta - I love this program SO much! If...
"topstyle 3.0 beta"
I'm very intrigued by the lately discussions about More Like This From Others. Its such an interesting concept, but I'm
Posted at December 17, 2002 03:24 AM
The Web Standards Project comes back with a new look and some great new information. Do you have a standards...
"WaSP update"
I am so excited to see this - TopStyle Pro 3.0 Beta - I love this program SO much! If...
"topstyle 3.0 beta"
I'm very intrigued by the lately discussions about More Like This From Others. Its such an interesting concept, but I'm
Posted at December 17, 2002 03:42 AM
The Web Standards Project comes back with a new look and some great new information. Do you have a standards...
"WaSP update"
I am so excited to see this - TopStyle Pro 3.0 Beta - I love this program SO much! If...
"topstyle 3.0 beta"
Ben Hammersley.com: Trackback, RDF, and the LazyWeb Here's the thing: I want to make a More Like This From Others
Posted at December 17, 2002 07:08 AM
The holidays are an exceptionally hectic time for my family and me. I'm already struggling to keep up with my mailing list and weblog reading. I was rather depressed I couldnÂ’t have been more active in the recent discussion of discovery file formats an...
Posted at December 17, 2002 08:02 PM
The holidays are an exceptionally hectic time for my family and me. I'm already struggling to keep up with my mailing list and weblog reading. I was rather depressed I couldnÂ’t have been more active in the recent discussion of discovery file formats and MLTFO (More Like This From Others) amongst others. Happy Holidays!
"Holiday Weblogging Blues."
I love it. Even in celebrating his birthday, Jon finds a way to expound on the virtues of the Web as an application interface. (Happy belated birthday Jon!)
"A Birthday Web Service from Jon Udell."
Mark Pilgrim has posted an amusing analysis of Dave Winer's recent comments on REST vs. SOAP.
"Analyze This."
The holidays are an exceptionally hectic time for my family and me. I'm already struggling to keep up with my mailing list and weblog reading. I was rather depressed I couldnÂ’t have been more active in the recent discussion of discovery file formats an...
Posted at December 17, 2002 08:08 PM
The holidays are an exceptionally hectic time for my family and me. I'm already struggling to keep up with my mailing list and weblog reading. I was rather depressed I couldnÂ’t have been more active in the recent discussion of discovery file formats and MLTFO (More Like This From Others) amongst others. Happy Holidays!
"Holiday Weblogging Blues."
I love it. Even in celebrating his birthday, Jon finds a way to expound on the virtues of the Web as an application interface. (Happy belated birthday Jon!)
"A Birthday Web Service from Jon Udell."
Mark Pilgrim has posted an amusing analysis of Dave Winer's recent comments on REST vs. SOAP.
"Analyze This."
Ben Hammersley challenges, More Like This From Others emerges. Ben Trott of Moveable Type fame releases mltfo.pl fufilling the dream.
Posted at December 18, 2002 10:39 AM
Geeks – at least those with any vision – dream of ‘the singularity’. For the social software geeks, this singularity
Posted at December 18, 2002 12:48 PM
Ben Hammersley quotes Blackbelt Jones: Matt 'BlackBelt' Jones has many good phrases, but his best is "LazyWeb": the idea that if you want to make something, just post about it and chances are people will do it first, or will have already done it. It wo...
Posted at December 19, 2002 02:38 PM
Ben Hammersley had an idea about relating entries based on trackbacks, and Ben wrote a module to make it work.
Posted at December 20, 2002 05:08 AM
Used on this site to show the outgoing ping information. 0.27 added MTPingedURLCount to show how many sites were pinged....
"PingedURL"
Ben Hammersley had an idea about relating entries based on trackbacks, and Ben wrote a module to make it work....
"More Like This"
Ben Hammersley had an idea about relating entries based on trackbacks, and Ben wrote a module to make it work.
Posted at December 20, 2002 05:09 AM
Used on this site to show the outgoing ping information. 0.27 added MTPingedURLCount to show how many sites were pinged....
"PingedURL"
Ben Hammersley had an idea about relating entries based on trackbacks, and Ben wrote a module to make it work....
"More Like This"
Ben Hammersley had an idea about relating entries based on trackbacks, and Ben wrote a module to make it work.
Posted at December 20, 2002 05:16 AM
Used on this site to show the outgoing ping information. 0.27 added MTPingedURLCount to show how many sites were pinged....
"PingedURL"
Ben Hammersley had an idea about relating entries based on trackbacks, and Ben wrote a module to make it work....
"More Like This"
Ben Hammersley had an idea about relating entries based on trackbacks, and Ben wrote a module to make it work.
Posted at December 21, 2002 01:13 AM
Used on this site to show the outgoing ping information. 0.27 added MTPingedURLCount to show how many sites were pinged....
"PingedURL"
Ben Hammersley had an idea about relating entries based on trackbacks, and Ben wrote a module to make it work....
"More Like This"
Ben Hammersley had an idea about relating entries based on trackbacks, and Ben wrote a module to make it work.
Posted at December 21, 2002 08:12 PM
Used on this site to show the outgoing ping information. 0.27 added MTPingedURLCount to show how many sites were pinged....
"PingedURL"
Ben Hammersley had an idea about relating entries based on trackbacks, and Ben wrote a module to make it work....
"More Like This"
Ben Hammersley had an idea about relating entries based on trackbacks, and Ben wrote a module to make it work.
Posted at December 21, 2002 08:13 PM
Used on this site to show the outgoing ping information. 0.27 added MTPingedURLCount to show how many sites were pinged....
"PingedURL"
Ben Hammersley had an idea about relating entries based on trackbacks, and Ben wrote a module to make it work....
"More Like This"
"until we have a universal taxonomy"
:-)
What about DAML, specifically the daml:equivalentTo
property? "When you say X, and I say Y, we actually are talking about the same category".It seems that this is intended to solve exactly this sort of problem ...
Posted by: DJ on December 11, 2002 01:12 PM
The DMOZ categories are I think the most sensible categorising you can get, and as dj says with some damn oily owls you could expose your internal mappings to this easily (well you could in RDF, RSS seems a strange constrained beast to me)
Then finding things in the same category would be easy once you'd scuttered up enough RDF. Like with foaf properties, this only really becomes interesting enough to make the tools once we've got enough marked up data.
Posted by: Jim Ley on December 11, 2002 03:41 PM
Sounds like fun - and I have scripts that do similar things (for example, my autoping.py Extracts useful information from a trackback.) Unfortunately, what I am really missing is how one makes the leap to index.rdf.
Ben, if you can provide a live test case and some prose that is closer to executable code (example: here is a URL, there is the trackback info, go to the URL it references, and there you will find...), I will be glad to provide you a something closer to a good starting point for the application you envision.
Posted by: Sam Ruby on December 11, 2002 11:23 PM
RDFQL is a great way of querying RDF documents. You can see it in operation with the RDF Query-O-Matic at http://weblog.burningbird.net/class_rdql/query.htm. I also have a Java version based on Jena, but no server for it at this time for you to try (will give you code if you want).
The key is what you want to do with it. Not sure from your writing. I'm assuming you want to get URL attached to the trackback entry, follow it back to the posting, use RDF Query to get the category from the MT-based embedded RDF, and then use RDF query on the index.rdf file to get URLs to items that are related to the person's category associated with yours. You then want to combine with yours for "Related items". Is that about right?
But where do you want to store this info? Do you want to create a new RDF/RSS file with entries to all related items for the post? This would be efficient -- you don't want to do this live or you'll end up being a 'bad neighbor'. With this, you can use a simple RSS display page to display the data when people click the button.
I wrote about this at my weblog (trackbacked as a matter of fact), but have to focus on RDF book right now. However, if something else doesn't pop up in the next few weeks, holler.
Sam, I don't suppose your Python script works with older Python does it? My ISP only supports Python 1.6.1 I believe it is. Would like to try your script.
Posted by: Burningbird on December 12, 2002 03:12 AM
Some URLS, even though Sam should be in bed by now, and Ben will be up again before Sam:
Start from the RSS feed for the pings for this entry, which you should probably assume the link in the HTML page will hand you, since it's oddly placed: take the TrackBack ID number (1017) off the TB URL, slap .xml on the end and tack it on the archive path to get http://www.benhammersley.com/archives/1017.xml
Parse out the item links, and go fetch
http://www.couchblog.de/couchblog/archives/2002/12/lazy_web.php
http://weblog.burningbird.net/fires/000731.htm and
http://philringnalda.com/archives/002407.phpRelease the TrackBack RDF from its HTML comment prison (boy, I wish I could have won that war), then from the rdf:Descriptions for your links (we all have individual entries, but you'll also have monthly archives with 30 or 50 separate sets of RDF, so you have to have the right rdf:abouts), extract the dc:subjects "Blogging", "Technology", and "rss/rdf".
Some handwaving and you've got the URLs for our RSS feeds...
Oh. Nico and Shell are nice, giving you a single RSS autodiscovery URL, but I gum things up with two. Maybe your first choice should be a link rel="home", hoping for a simpler RSS autodiscovery situation there? ...
And then you fetch those RSS feeds, and parse out any items with the same dc:subject as the item that sent the ping, and Bob's your uncle.
I really like getting to read the unexpected, so the sort of results it will get with the current category situation will work great for me, but in general I think it would work a lot better if MT uncoupled the archive category from the metadata categor(y|ies), and made multiple categories easier. Not that I know quite how to do it, but right now having a good fine-grained set of categories also means having a huge bunch of archive files with two or three posts in each, making them hardly any better than keyword searching. If you could define a set of broad archive categories, and then assign an entry to a narrower subset as its primary metadata category, you could probably do better stuff with the results.
Posted by: Phil Ringnalda on December 12, 2002 05:24 AM
Shell and Phil have it spot on, but a few things might help.
Firstly, by default the rdf page for MT is /index.rdf - even if there is no link rel, this is worth a try first.
Second, yes, Bb, converting it at the end to a static RDF file is the idea. I was thinking this script could sit at the end of the receive-trackback function. The trackback comes in, the script runs, builds a static file, and that's it. The page is a kind of 'snapshot' of the state of the conversation at that time, which is actually more interesting in my mind than a dynamically building page.The reasoning behind doing it this way is that the very act of trackbacking/linking implies that two entries are on the same subject. The author is basically saying that the two taxonomies touch at this point.
Of course, they probably diverge with other entries (some people's categories are massive supersets, others are very precise indeed) but this just adds to the serendipity. However, this divergence makes it hard to use DJ's DAML syntax above, so we have to take each entry on its own.
Of course, once we've done that, we can move onto pulling URLs out of referrer logs...
Posted by: Ben Hammersley on December 12, 2002 09:37 AM
Having all sorts of fun... ;-)
In Shelley's page, dc:subject is an element, in everyone else's, it is an attribute.
Python's SGML parser doesn't know what to do with Phil's CDATA wrapped scripts.
Shelley's pages for individual entries don't have application+rss links.
Default to /index.rdf sounds like a good idea, but in practice http://www.couchblog.de/couchblog/archives/2002/12/lazy_web.php actually maps to http://www.couchblog.de/couchblog/index.rdf, so an interative search across up the directory hierarchy is required.
Posted by: Sam Ruby on December 12, 2002 02:00 PM
OK, it looks like Ben's RSS feed isn't valid, so I will stop short of that part of the exercise.
Code implemented to date can be found at mltfo.py. It does very little error recovery as I find the Python stack tracebacks helpful during debugging. It is also designed to handle the test data provided.... given what I have seen so far, other test data would likely require more support to be added.
Example output:
python mltfo.py http://www.benhammersley.com/archives/1017.xmlhttp://www.benhammersley.com/archives/003371.html
Content Syndication with RSS
http://www.benhammersley.com/index.rdfhttp://philringnalda.com/archives/002407.php
rss/rdf
http://www.philringnalda.com/index.rdfhttp://weblog.burningbird.net/fires/000731.htm
Technology
http://weblog.burningbird.net/index.rdfhttp://www.couchblog.de/couchblog/archives/2002/12/lazy_web.php
Blogging
http://www.couchblog.de/couchblog/index.rdfEnjoy!
Posted by: Sam Ruby on December 12, 2002 10:30 PM
Ah, Sam - sweet.
It's bed time for me, but I've fixed the feed issue...
Posted by: Ben Hammersley on December 12, 2002 11:47 PM
How about ISBN instead of DMOZ?
Posted by: GilbertZ on December 13, 2002 01:43 AM
Sam -- Python version?Also, hehe, I created an RDF/RSS file that validates in the validator, but isn't valid RDF/RSS according to the specs. SORSS.
Posted by: Burningbird on December 13, 2002 11:35 PM
Hi,
Here's an implementation of "More Like This From Others" in Perl. Another success story for the LazyWeb! :)
Posted by: Benjamin Trott on December 13, 2002 11:55 PM
what about the work being done by the WayPath project? http://www.waypath.com/
Posted by: paul on December 16, 2002 07:49 PM
Done! Implemented! Nice!
Posted by: Ben Hammersley on December 16, 2002 07:55 PM
I have to agree with paul, though I'm a little biased. The Waypath Project lets you add a one-click link to more-like-this results to every post, and it has the tools to let you embed more-like-this results in you templates, formatted however you want (via css).
Because it doesn't rely on trackback (or pingback), external taxonomies, or classification mapping, it immediately opens up your more-like-this world to people you don't know and who may not be using the same classifiers you are.
And while there are still a lot of features we're working on, the basic service is in place (and being used by growing ranks of bloggers). So, go ahead, be as lazy as you want.
Posted by: steve on December 16, 2002 08:12 PM
Take a look at the link above, if you like: what I'd like to see, in a larger context that weblogs, is some way to find "more like this" no matter what I'm reading. There are obviously some Google-sized scale issues there, but more to the point, what do you mean when you say "more like this?" Are you relying on the content creator to determine what the content is about? Look at the links Waypath found to the link above (tabbed browsing is great for stuff like this): they're about document management, even though the words "document management" never appear in the source document, the "this."
I would argue against any system that requires the use of a taxonomy or user-input: given the horsepower we have available today and the deper understanding of how to model content, we should accept nothing less than context-based discovery/detection.
Posted by: paul on December 17, 2002 04:43 AM
Sorry, the page I reference is linked from my name . . . .
Posted by: paul beard on December 17, 2002 04:46 AM
So, what I want is a little app that takes the trackback. Follows it back to the originating site, find the RDF snippet, takes the index.rdf, and gives back all the entries within the index.rdf that are on the same subject as the trackback one.
I think the bones of this are in place at Mark Pilgrim's diveintomark site: his "further reading" contraption does the "pull referrers out the access log and pull the referential text from that site for inclusion in the referred-to page."
Got all those references?
Posted by: paul on December 17, 2002 04:51 AM
Unfortunately, my experience with Waypath tends to be more like your mozilla building entry, where you get one link that doesn't seem to have even the tiniest relevance. Though MLTFO is likely to be a bit rough, too, when a very specific entry gets pinged by an entry in a broad category.
Posted by: Phil Ringnalda on December 17, 2002 05:21 AM
I agree that WP is often wide of the mark, but I'm more interested in something that fails with the right intentions. I don't see how MLTFO enables serendipity: it seems to be finding weblog entries that already link to "this." I like finding what I want without knowing I was looking for.
MLTFO seems too much like "further reading" and we already have that. As much fun as it is to trawl through the referers (I've beguiled away many an hour doing just that), I'm more interested in people who aren't reading what I'm writing but are writing similar things. Where MLTFO seeks to strengthen existing links, I want to forge new ones.
On WP's occasionally lousy results, the shorter the source article, the less likely you'll get good results. It needs context and the more the better.
Posted by: paul on December 17, 2002 07:03 AM
Would XFML be useful? (See http://www.xfml.net/ and http://www.xfml.org/ . Someone has even written a MT template for producing an XFML feed: http://poorbuthappy.com/ease/000494.html.
Posted by: Howard Fore on December 18, 2002 11:16 PM
Would XFML be useful?
Looks interesting but it seems to require some assembly: I'd be more excited about something that didn't require me to define a category for a given article. I find myself wanting to define new categories for stuff I've written -- as if I had any idea I'd get to 600+ entries, never mind the variety.
I'm reluctant to quote myself, but . . . .
But what's really needed, I think, is the next step after assigning or deriving a vector. Look at the old Dewey Decimal system or the new CIP structure: why can't that be done for documents, such that each document has a unique code based on what it is/contains?
I still think there's some kind of hierarchy or structure waiting to be built. It seems likely to me that it will have to be based on a lexicon or domain-specific thesaurus: one for law would not work for aerospace engineering, frinstance. So there's no way to auto-generate the classification map: I do see that as the end goal. I'd like to think that, given a meaningful sized document and therefore a representative extract -- this starts to resemble statistics -- that a given document could be "filed" by some code or other arbitrary representation and then evaluated against others.
Posted by: paul on December 19, 2002 04:39 AM