Attention, Reading Lists, RSS, etc, etc

Attention continues to get my attention. David Sifry, in his recent update on the blogosphere and its staggering growth, says:

We track about 1.2 Million posts each day, which means that there are about 50,000 posts each hour. At that rate, it is literally impossible to read everything that is relevant to an issue or subject, and a new challenge has presented itself - how to make sense out of this monstrous conversation, and how to find the most interesting and authoritative information out there.

Alex Barnett posted on this issue:

The live web discovery problem is different type of discovery problem to that the traditional search engine space has been trying to solve. Companies such as Technorati, Icerocket, PubSub, Memeorandum, Tailrank, Digg, FeedDemon, Rojo, and Bloglines and many other start ups that have cropped up in the last couple years recognize this and are helping us navigate the torrent. However, in my view, what's missing from the current generation of the aggregators, feedreaders and live web discovery engines is the ability to scope these services against my attention data. Some of these services provide tag and keyword RSS search subscriptions and have some personalization features.  These are steps in the right direction, but we've got a long way to go.

(There's a podcast available with Alex, Joshua Porter and Steve Gillmor discussing attention.) Which makes me recall Herbert Simon's words:

What information consumes is rather obvious: it consumes the attention of its recipients. Hence a wealth of information creates a poverty of attention, and a need to allocate that attention efficiently among the overabundance of information sources that might consume it.

Reading Lists, as I've blogged before, are hot and are being talked about in the context of an 'attention-based recommendation system'. (Listen to another of Alex's podcasts, here, with a discussion between Alex, Danny Ayers, Joshua Porter and Adam Green about Reading Lists.) Dave Winer's guidelines are lucid and helpful (and see his OPML Editor doc) and I have really enjoyed Danny Ayers' take and this comment by Darren Chamberlain:

I think I don’t get the idea of a reading list. Is it just the portion of a blogroll that you’ve been reading most recently (the blogroll’s intersection with your attention data)?

dd's comment points to a key significance of Reading Lists, their dynamic nature. EirePreneur has a post touching on Reading Lists but focusing on Feed Grazing and (wait for it) Web 3.0 that has set me thinking, and Danny Ayers' comment there ('the near-future of the web is going to be a generalisation from a Web of Documents to a Web of Data') is my excuse for not yet blogging about last week's conferences.

Alerted by Alex and Robert (and Greg Linden's comments on the latter), I'm playing with Megite (my personal Reading List here; not a good idea, it seems, for me to have included the BBC News feed — it swamps everything), and have now also gone back to Findory and TailRank. (There's a post about all this by Richard MacManus, too.) Alex:

Megite is going letting me do what I've been asking Memeorandum (or anyone else that will listen) to let me do for ages  - to pivot off my own OPML file. The feature isn't switched on for everyone yet, but I've pinged the Megite developer, Mathew Chen, so hope to hear from him soon. … I'm more convinced than ever that the ability to render a personalized experience based on Attention data is where its at. And I'm not talking about just clickstreams. Your OPML file (specifically your list of RSS subscriptions) is one example of this Attention data set. It says a lot about you: the topics your interested in and the people you listen to, and much more. There is plenty more Attention data that can be leveraged though. My tags, my wishlist, the books I own, etc. We're just at the beginning of the Attention Engine race.

In the comments to Alex's post, Greg Linden says: 'Thanks for trying Findory! The relevance rank is not random nor is it solely based on your OPML file.  Findory decides what is relevant based on the articles you read. Play with it, click a few articles, and watch how it focuses in. Findory learns very quickly'. And Kevin Burton: 'TailRank has had this live for 2 weeks now'.

Swarming media has a post on some other, related implications of all this — the way we're projecting our deterritorialized, multiple identities in cyberspace:

The obvious unwanted social implications extend to surveillance and impersonation, but culturally, we are creating selves outside ourselves. Many-tendriled projections.

Compare James Governor on Declarative Living.

Meanwhile, FeedBurner's FeedFlare API (the release of which coincided with last week's Future of Web Apps conference) has got my attention:

The really big idea … was … the notion of providing a universal framework/API to enable any third-party web service to integrate with a publisher's content, without concern over what content management system the publisher is using.

The 101 ideas FeedBurner published for FeedFlare underscore the role of RSS as a way of gluing things together. Mitch Ratcliffe on the original FeedFlare announcement:

Using metadata this way will allow greater integration of intelligence in the management of feeds. The announcement talks about more browser-friendliness, which is a big plus, but RSS is fading into the communications between applications and, I think, that's where it will take deepest root.

Kevin Burton:

(FeedFlare) should allow more innovation in the space.  For example I could add TailRank features directly in FeedBurner.  Other smaller companies could add plugins for their content as well.

In a year or two, what will be the place and nature of RSS aggregators and these rich RSS feeds? Richard MacManus has a post today declaiming, 'Personalization + Clustering is the next big step in RSS. If 2005 was about Aggregation, then 2006 is all about Filtering.' Danny Ayers focuses on the technicalities behind this and in the comments adds: 'the smart aggregator (with hooks into things like the Technorati API and a bit of P2P) is probably a quicker route than trying to put all the processing online'.

Union Square Ventures invested in FeedBurner believing that RSS will become mainstream, but they, like Fred Wilson, know there's some way to go yet. Matt McAlister's gloomier still. Me? — I think Lloyd Shepherd has it right: 'the fact is that RSS is gluing all sorts of things together at the front end and the back end. … it’s entirely understandable that the RSS front end is still a bit squishy and unfriendly - people are still trying to get to grips with the possibilities of it at the back-end. Not because people are stupid, but because those possibilities are just so huge'.

Back to Attention. The Guardian picked up on this last week and advertised AttentionTrust.org. I joined this a while back and am now beginning to see its value through using the AttentionTrust approved service, Root Vaults. You can download AttentionTrust's Attention Recorder extension for Firefox here and you have the option either to record your attention data direct to your hard drive, or to Root Vaults or ACME Attention Service.

These are some of the things to do with attention, RSS, etc that have been crossing my radar recently. (There are others, but I'm sticking here with the ones that have really preoccupied me. Companies like Attensa are on my screen, too …)

Technorati tags: , , , , , , , ,

Catching up with Reading Lists

OPML. OPML. OPML. How could I forget thee?

Reading lists are OPML documents that point to RSS feeds, like most of the OPML documents you find, but instead of subscribing to each feed in the document, the reader or aggregator subscribes to the OPML document itself. When the author of the OPML document adds a feed, the aggregator automatically checks that feed in its next scan, and (key point) when a feed is removed, the aggregator no longer checks that feed. THe editor of the OPML file can update all the subscribers by updating the OPML file. Think of it as sort of a mutual fund for subscriptions.

    OPML is a really useful file structure that just about everyone who uses a feed aggregator, like bloglines, is already using without necessarily knowing it. Most readers keep subscribed feeds for a user in OPML format, for easy importing and exporting. If you export your OPML feed you get a XML file of your feeds, which other feed readers understand.

    The problem with OPML files from readers is that they are static, meaning I can give you my OPML file but you will never know if I add or delete feeds unless I tell you and give you the new file. All you get is a snapshot of my feeds from the moment that I share my file with you. Dave [Winer] thinks these files should be dynamic, which means that I can share my opml file, or as he calls it my reading list, and anyone who subscribes to it will always have the current version, no matter how often I amend that list. There is very little technology needed to allow this to happen - the various feed readers simply need to agree to support dynamic lists and allow people to share them permanently. Dave’s trying to make this happen. If he succeeds, we’ll all be able to subscribe to reading lists from people we trust on a given subject, and good feeds will be that much easier to find. … In a comment, Eric Lin writes:

    I could easily see this not only as a way to share my reading list with others I know, but also to be matched with others I don't know with common interests. What if the system could match me with other people who have similar tech, music or lifestyle feeds as I do. It would be a fantastic way to make new connections as well as strengthen existing ones, and I could see communities forming around overlapping feeds. These communities might be stronger than those that form around a single website because they'd have more in common.

  • Nick Bradbury: Reading Lists for RSS
  • In a nutshell, the idea is that you'd subscribe to an OPML document which contains a list of feeds that someone is reading, some organization is recommending, or some service has generated (such as "Top 100" list). Changes to the source OPML document would be synchronized, so that you're automatically subscribed to feeds added to the reading list. Likewise, you'd be unsubscribed from feeds removed from the original OPML.

(Thoroughly indebted to Alex's post, Reading Lists = the killer app for OPML.)

The social aspect of OPML, 'communities forming around overlapping feeds', is really interesting.

Technorati tags:

clevercactus feedexplorer

'feedexplorer is a simple, free application that allows you to navigate the subscription lists published in the Share Your OPML site, discover new feeds, choose among them, and generate subscription lists to use in your news aggregator. Other functions, such as search and sorting by number of subscriptions are also provided. It runs on Windows, Mac OS X, Linux, and other systems. '

Nick Bradbury interviewed

'Every year we hear about the imminent death of packaged desktop software. Still, you appear to be gainfully employed. Will we be reading blogs in desktop apps three years from now?'

Nick Bradbury: 'Rumors of the death of desktop applications are greatly exaggerated :) The line between the desktop and the web will continue to blur, but for the moment desktop apps simply provide more of what customers want.' Read more: since1968.

Oskar van Rijswijk & FeedDemon

logos on search channels:

One of the handy features of the RSS feedreader FeedDemon are the Seach Channels. You can make a feed - directly in the FeedDemon UI - from a keyword search in the great RSS searchengines Feedster and Daypop. But more can be added with XML-files. You can download an extra set on my FeedDemon page, at the bottom of the page. It’s a zipped directory with Search Channels for: BlogDigger, Daypop, FastBuzz, Feedster, GoogleNews, RSS-Verzeichnis, WayPath and Yahoo!News. Just copy this directory to your FeedDemon > Data > search directory on your PC and you are ready to tango. And you know: you can also search for hundreds of RSS feeds on my RSSlinks page and or synchronize them on my OPML Directory page.

Salute to Dave Winer: reading lists

Dave Winer's initiative at 'Share Your OPML' is thoroughly inspiring in its generosity and imaginative scope:

One of the innovations flowing out the Share Your OPML site is the idea of reading lists. An expert in a given area puts together a set of feeds that you would subscribe to if you want a balanced flow of information on his or her topic of expertise. You let the expert subscribe to feeds on your behalf. I've gotten the first taste of what this is like by reading the aggregator page on the Share Your OPML site. As new sites come on the Top-100, as the aggregated interests of the community shift, I automatically start reading sites I wasn't reading before. I don't have to do anything. I like this. So at last Thursday's Berkman meeting I asked two of our regulars, Rick Heller and Jay McCarthy, to start doing these reading lists, and Rick is ready with what he calls a list of "political blogs that provide a balanced diet of liberal and conservative views." Now I have more work to do, to create a user interface that lets Rick edit his list at will, and presents an easy way for you to subscribe to his list so he can automatically subscribe you to new feeds (and unsubscribe you from others). The technology is not that hard, but it's essential, imho. Two comments. 1. I'm talking with other developers about building around this idea, so there will be another round of open formats and protocols building on RSS, OPML and XML-RPC; and 2. No patents.
Scripting News.