Alex Barnett's blog : Attention

Moving my blog

alexbarn — Sat, 02 Sep 2006 21:38:00 GMT

OK, so I moved my new Alex Barnett blog to here for a number of reasons, explained here at my, er, new blog.

My musical attention (data)

alexbarn — Sun, 25 Jun 2006 03:56:00 GMT

Earlier this month I mentioned that at the Content 2.0 event I asked LastFMs product manager, Matthew Ogle about the ability to export my data out of the system so it could be plugged into another. After all, the music I listen to, for how long and when I listen to it, user-defined tag and ratings, is my data - my musical attention data.

Matthew responded with a resounding 'we're planning to - absolutely'. I asked what format they would use and there Matthew admitted he wasn't so sure. An open standard, yes, but which? Is there one than can carry the data? Or does one need to be created? Offline, we discussed a few options.

Now, apart from the principle that my data should be something I can use as I like (meaning I should be able to export it out of the system that was used to generate that data in a non-proprietary format), there are other reasons for why you might want to 'own' your musical attention data.

At the basic level, my music library needs to be expressed in a way that allow me share with other people and systems (I'm not talking about the actual music file here, just the metadata). Today you can export your Window Media Library info from Windows Media Player (you need this plug in - it works with WMP 10 and 11 beta), but there is no music library standard, let alone one that can include other attention data.

Now, there are few playlist formats (including open standard formats - W3C's SMIL and Lucas Gonze's XSFP), but a playlist format is not a music library format. Playlists are collections of tracks.

Mike Torres proposed a good idea, one that comes at this from the perspective of music subscription services. The problem that Mike highlighted is that once you've stopped paying the subscriptions for the music your preferences (musical attention data) are lost along with the access to the music.

"Think about it; it isn't the actual bits on your hard drive that you care about - it's the fact that you spent 43 hours last month searching, previewing, and selecting that music. It's the metadata associated with the songs that's important; the actual WMA files aren't where the value is. You don't want to repeat the effort. Why treat your music subscriptions any differently from your RSS subscriptions? Those are portable. Your email archives and bookmarks are portable. Your music subscriptions should be too.
"

So there are two things here:

1. An open standard format to describe a music library. Scenarios it would support include:

Exporting and Importing - taking my music library out of a system to be able to plug into another (this is the music I like)

Sharing - publishing the library on the web (not the actual music, but what music)

2. An open standard format to describe the attention data about the music (how often I listen to which tracks, when I listen to them, how I've tagged them). Scenarios it would support include:

Recommendation - plugging in my musical attention data to get recommendations

Mike goes on to propose OPML may be a good format for the cataloging requirement.

"just let people export and import their subscription libraries natively via OPML."

Yup. Tom Morris had the same idea So did Kosso (OPML Listening Lists) and others.. In February, I even made listening list out of OPML. In fact, given that Podcasts are delivered via RSS enclosures, it's no surprise to see podcatching services have been using OPML as a 'directory' formats- import and export away. Some are share them while others are hacking to force the function.

Thinking out loud now, where do we go from here? Well, on #1 above, back to the Window Media Library export. It exports to a number of different formats, including XML. Maybe it could export OPML? Or someone could build a plug-in to transform on the fly? I don't know, but in this case it doesn't seem too far to get there.

And on #2 (the musical attention data format) - couldn't the attention.xml format (more on this here) be used as the basis for a musicalattention.xml? Sorry, just crazy thoughts...

Coping with Information Overload (video)

alexbarn — Fri, 09 Jun 2006 17:19:00 GMT

About four weeks ago I made my way over to ZDNet's studios in San Francisco to be interviewed by Mike Elgan for DevSource and discuss a number of topics close to my heart. The video (15-ish mins) is now published.

Mike asked me about microformats, tagging, information overload, OPML, attention data and of course web 2.0. From the article introducing the interview:

"Are your users drowning in data, and you have no idea how to help them navigate through the information available? It doesn't help to know that you're the only one without a life preserver. Fortunately, the Web is evolving, and new tools are being created to help you — and the people you serve — navigate the expanding information space.

...In our latest DevSource video, Barnett shares his expertise on Web 2.0, OPML, tagging, data programmability, and microformats."

I tried to stay away from pure geek talk and instead talk about some of the benefits and the application of the various technologies we discussed. I touched on Dave Winer's Share Your OPML (see my post on it here), Greg Linden's Findory, the Del.icio.us bookmarking service, Josh Porter's 'the Del.icio.us lesson', and enterprise tagging.

Thanks to Mike Elgan for thinking of me and asking such well researched questions :-)

The interview is part of the DevSource video interview series. Other interviews worth checking out are with Rob Howard talking about the evolution of ASP.NET, Jakob Nielsen discussing usability, Jesse James Garrett (who coined the 'AJAX') on AJAX and the Jeremy Bailenson interview about virtual reality.

Tags: web 2.0, tagging, microformats, attention, opml, video, interview

Share Your OPML (it had to happen)

alexbarn — Mon, 08 May 2006 09:08:00 GMT

So, who would have thought that OPML would be a big deal? ;-)

The interest in Dave Winer's latest project, Share Your OPML, is the reason for OPML's Attention this weekend. And with good reason too. Dave hinted at this project at last week's Seattle Mind Camp where we discussed Attention. This app / service is so inline with the OPML as an Attention data file thinking.

I've shared mine here. The service doesn't offiicially launch until Monday, but there's already plenty of OPML sharing going on. Steve Rubel, Robert Scoble, Michael Arrington and JWynia are blogging (and sharing) this too, and plenty more.

Nice thing is I can see who is sub'd to me. So far on this two feeds (one gen'd by this blogware, the other is a Feedbuner feed I created in September.

Who's Sub'd to Me...

Alex Barnett blog (feedburner - this one's better ;-)

Les Bain	118
Alex Barnett	593
Rod Begbie	232
couchblog	274
Peter Giger	49
Bela Labovitch	100
Tom Morris	334
Chris Pirillo	1441
Robert Scoble	99
TibsBits	102
Kate Trgovac	134
John Tropea	53

Alex Barnett blog (the other feed)

michael arrington	373
Alex Barnett	593
Christian Cadeo	310
Darkwookiee	154
Brian Johnson	96
paulbwalker	119
Murray Robinson	20
cori schlegel	402
J Wynia	457

Most prolific subscribers:

Here's the top ten as they stand right now (how many feeds they sub to)

1.	Chris Pirillo	1441
2.	enoch choi	960
3.	Larry Borsato	678
4.	Douglas Sorocco	638
5.	flycook	636
6.	Alex Barnett	593
7.	Cyb Alive	519
8.	Randy Charles Morin	486
9.	JDo	476
10.	J Wynia	457

-
Update: I just realized there was a 'show me people who have subsciptions like mine' thing. Very powerful affinities: Here's mine (top ten most like my OPML file)

			Strength
1.	Pete Gilbert	548	19.667
2.	EirePreneur	45	11.501
3.	Randy Holloway	219	6.052
4.	Robert Scoble	99	6.012
5.	Jeff Clavier	289	5.337
6.	Peter Giger	49	4.833
7.	Dave Winer	414	4.728
8.	John Tropea	53	4.663
9.	Richard Hamilton	192	4.632
10.	Eddie Dickey	151	4.613

-
Tags: OPML Attention RSS

it's about throwing stuff away

alexbarn — Thu, 04 May 2006 19:20:00 GMT

I had the pleasure of meeting with Steve Gillmor in San Francisco earlier this week. We met downtown where he picked me up and drove me to a cafe in a suburb - I don't know where - and we chatted over a late lunch and lattes.

We discussed the past and future of the internet. It was wonderful opportunity for me ask the Attention Tsar loads of dumb questions (my manager told me I'm good at that) and to get deep into the topics of common interest to us.

There was a recurring theme to our discussion - Steve pretty much summed up it up in a comment he made yesterday :

"Search is dead. The amount of viewing/listening time is finite. It's not about finding stuff; it's about throwing stuff away."

Tags: Attention

Attention Clouds

alexbarn — Thu, 04 May 2006 08:29:00 GMT

Here's an attention cloud for you.

az.

"The Intention Cloud is a web application supported by Perl and AJAX technologies that gathers and displays the queries performed on search engines by their users in an intuitive output using weighted lists, also known as tag clouds."

Tags: attention tags tagging

Attention 2.0

alexbarn — Mon, 01 May 2006 08:29:00 GMT

(I had to call this post something and given this pic, I couldn't resist...)

(er, that's Attention two Dot oh, folks...)

The Seattle Mind Camp 2.0 Sunday session on RSS and OPML led by Dave Winer was a fascinating hour, spent discussing, well, RSS and OPML and more. In the room were around 50 geeks, some with little or no knowledge of OPML (all knew of and or used RSS). It was an intimate affair.

A great deal of the discussion revolved around the Attention, often explicitly. We kept going there, it just gravitated naturally in that direction. What struck me this morning is how there appears to be great deal of common thinking around the Attention topic. Sooo much to do and sooo many opportunities before us. Now, I realize this was an uber-geeky crowd but they were very receptive to the cries of 'it's my data!!'. Edge cases, or early adopter? Who knows

It's a shame Dave wasn't able to attend the session I ran last night entitled Attention and My Data where for the time I gave a talk based my writings and observations (and podcasts) on the subject over the last year and a half or so. I spoke of Goldhaber's 1997 essay, Information overload, Attention data, the evangelizing of Attention by Steve Gilmor, the Attention trust principles, Attention Engines and Attention Lenses, OPML as Attention Data and the notion of 'your Attention Data being yours' and my 20 thoughts on Attention. From my perspective, it was useful exercise. I tried to tie a number of related strands together into one coherent story, one that I hopefully will have nicely polished by the time the Content 2.0 event happens in London this June 6 (I'm booked to do 15 mins preso on Attention there toward the end of the day).

Like I said, Dave Winer's the session was intimate - the crowd was small in numbers but smart and included the likes of John Musser who runs ProgrammableWeb and Sean Lyndersay, a Lead PM for RSS at Microsoft,

(John and Sean )

(other smart geeks such as Jack William Bell, Dennis Hamilton (aka Orcmid )

But of course, this has to be my favourite (English spelling) pic of my Seattle Mind 2.0 Flickr set:

>>An aside. Yesterday was the first time I met Dave and today was the first time I heard him in the real-world talk about his perspective on the potential of OPML. There is something about the 'real world' level communication the seems to ensure that what is being said is more penetrative to the ear and mind. I'm not talking about the interaction side here. I'm talking about the listening bit. It's to do with the focus I have as a listener I think - the words can be the same when I listen to a podcast or a watch an video interview, or read the written form yet the 'liveness' provides a whole other level of immersion for me. I had the same sense recently at Mix06. Listening to Tim O'Reilly up close and in person seemed to make sure I really tuned in, and so 'grokked' more of what he said, even though the ideas themselves weren't new to me. Today, I grokked more.<<

Up early tomorrow AM, I'm off for a secret mission in Fran CanDisco. Will blog later...(not very secret)

My Attention writings

Tags: Attention, OPML, MyData, MindCamp 2.0, Web 2.0

One Big Conspiracy

alexbarn — Mon, 24 Apr 2006 07:39:00 GMT

According to Charlie Brooker, it's all one big attention conspiracy:

"Install this driver. Now update it. Now update it again. Register to login to our website. Then validate your membership. Forgot your password? Click here. Now there. Fill out this form. And this one. And this one. Please wait while TimeJettison Pro examines your system. Download latest patch file. Please wait while patch file examines own navel. Remove cable. Insert cable. Gently tease USB port with cable. Yeah, that's it baby. That's the way. Now show us your bum or I'm deleting your inbox."

-
Tags: Attention

Social Search

alexbarn — Fri, 14 Apr 2006 19:50:00 GMT

Business Week has an article outlining some of the 'social search' developments underway at Microsoft. What is 'social searching'?:

"Microsoft plans to unveil a question-and-answer social-search tool in the coming months, says Justin Osmer, senior product manager for MSN. The feature will let users direct questions to a specific universe, such as a group of friends, rather than to get automated lists of results from a generic search engine"

In the same article, it hints at similar efforts by Yahoo and Google but points out that the advantage that Microsoft has over competitors is the 205 million IM and 230 active Hotmail users.

(I should point out my usual disclaimer at this point and explain that I'm not privvy to any of Microsoft's social search product plans at this stage).

It's an interesting direction and one that follows on from the 'using your social network as a filter' idea. Today, the 'live' internet of content is filtered via my OPML file (and anyone esle that is using an RSS reader). The blogs posts I read, articles pointed out and discussed, photos and music suggested happens mainly via my OPML file. My OPML file is one cut of my various social networks. Extending this to the 'liveness' of knowledge in the form of instant answers via your social network looks set to be a major area of innovation over the next couple of years.

Update: Memorandum and Megite have some good links to others' thoughts on the story.

Tags: Attention, social software, social networks, search, Microsoft, OPML

GestureBank Recorder

alexbarn — Fri, 14 Apr 2006 07:41:00 GMT

Steve Gillmor provides some details on how the existing Attention Trust recorder (ATX) relates to GestureBank - the codebase is forking. GTX will be the GestureBank Recorder.

Sounds like an IE version of the ATX and GTX recorders are on their way too.

The beta's nearly ready...

"Now, with GestureBank and its open pool of metadata, we have the opportunity to take the next step toward creating value and ratifying the Trust's other principles–property, mobility, and economy. See you at the Pool."

Tags: Attention

Information overload in 1848

alexbarn — Fri, 07 Apr 2006 08:16:00 GMT

Information overload in 1848:

"In 1848 the New York businessman W.E. Dodge was already feeling the effects of always-on connectivity:“The merchant goes home after a day of hard work and excitement to a late dinner, trying amid the family circle to forget business, when he is interrupted by a telegram from London.” "

John Udel's review of the Victorian Internet.

Tags: Attention

'myware' but without the 'my'?

alexbarn — Wed, 05 Apr 2006 04:41:00 GMT

I have to agree with Steve Gillmor on this:

"All proprietary clouds of data obtained without the users' permission are tainted, and it's my bet (GestureBank) that given a choice between an open pool and any proprietary one, the open pool will implicitly be more trusted."

Not surprising that I would agree, I suppose. But why?

Well, as Kevin Burton of Tailrank points out the sp*ware companies are getting into the Attention act:

"Looks like Claria wants to enter the Memetracker (or customized home page) market by shipping a browser based toolbar."

So here's the thing. In this discussion with Kevin we both concluded that the ability to 'own' (manage, edit, purge, delete, share, not share, etc) *your* attention data is the right way to go (well, we've agreed on this ages ago, but we agreed again...).

On a parallel track, the observation Kevin makes of Findory is a good one - the 'implicit' nature of *some* what's going on. Findory's recommendations algorithms are very good indeed. But, from my perspective, it is has a great starting point: a subset of my attention data - my OPML file. From there it uses my clicktrack behavior on the Findory site to recommend content to me. A very rich pool of content - I've already told it what interests me via the OPML file I pointed it to. To stress - in the case of Findory the clicktrack recording is limited to my behavior on Findory. That's it. And that's ok. (Findory's developer Greg Linden might point out at this stage that the OPML file is not necessary to provide a very relevant experience - just a few clicks in one of the pre-defined channel such as 'technology' would do it. I'd believe him too - he's the guy who built Amazon's recommendation engine that's made the company a gazzillion dollars. My point here though is that it is my OPML - the people I want to read - are my uber-smart attention filters. They write about and point to the stuff I'm really, really interested in and Findory acts as another layer of value on top of that.)

Claria's proposition is that it'll track you everywhere - 'myware' but without the 'my'. Can you edit the data it tracks? Can you export it? Can you expose it to another service and treat the data it collects as your data? Is it your data? Nope, nope, nope, nope and er, nope. Sounds great doesn't it?! Install software that tracks your everyclick so marketers can spam you! Lovely! Sign me up!! Not.

This is where the Attention recorders come in (as Jon Galloway pointed out in the comments to my related post). This is where Steve is coming from. There are others thinking this way too. The questions I have are:

which of the Attention recorders will play an open Attention data game?
by open I mean - which attention recorders will allow the users' data to be edited, viewed, shared, purged, exported and controlled - fully - by the user?
and which of these recorders will act as a service that others services (or people) can connect to (or locked out) with the user's permission?
and...which developers of the newstrackers, aggregators, attention engines and raft of other services that would benefit in accessing that data with the user's permission will be smart enough to see the opportunity here by allowing the third party, *trusted* recorder-provisioned, user-controlled attention data pools to plug in?

I've still not got the GestureBank invite. But from everything Steve has written about his idea it looks like its going to set the standard in the open myware / mydata / attention data / recorder / connector scene.

Tags: Attention, mydata, myware, OPML, web, tech, web 2.0

Your browser history as Attention data?

alexbarn — Sat, 01 Apr 2006 19:17:00 GMT

Have you thought of your browser history as attention data? I have some thoughts on this. Not all good.

Two companies are thinking this way. Tailrank is one. You can go to this Import page and give the site permission to look at your browser history for blogs you've visited and make browsing recommendations based on that data.

The other is company is Brightcove, Jemery Allaire's online video start up. In this post it mentions that a forthcoming release will "prebuild the recommendations from your browser history using some crazy AI kung-fu.". (Update - apparently this Brightcove post was an April fools joke. It got me....but as you can see - entirely plausable....)

Your browser history as attention data?

There is good, bad and ugly in this....

The Good

In Tailrank's case, you need to give the site explicit permission for it to trawl you browser history. That's a good thing.

Your browsing history is valuable. Very valuable. Cool stuff can be done with it.

But giving up your browser history isn't just a case of having a cookie placed so it can track your behavior on the cookie issuing site. We're talking about all your browsing history (i.e. the urls you've visited down to individual page level) across all sites. If the value proposition is right, this sounds like a good deal - I give you my browser history and you provide personalized content / recommendations. This may be appealing to some, but not me.

The Bad

The implicit bug in the 'reading the browser history' approach is the realization (for me at least) that a malicious site could do something you don't want it to do - i.e. look at your entire browser history (up to when you last deleted it). Why does this matter? Well, apart from the privacy invasion, there are real security concerns here.

Before I go further, I need to point out that you need to give explicit permission for the Tailrank site to be able to look at your browser history under default security settings of most browsers. You can't just land on a page for a site to look at your browser history. You need to click on something that gives the site permission to look. This is true for IE and Firefox. I tested Tailranks 'Auto-Configure' implementation (uses javascript). In IE at 'Medium' security, after I click the 'Auto'Configure' button Tailrank chugs away and trawls just fine. At 'High' security setting IE, Tailrank barfs - the javascript is disabled.

The problem is that you can be fooled by malicious sites into giving this permission. You could be fooled into clicking a link or button that doesn't do what is says it's going to do and does something else instead. That's bad.

And The Ugly

What's the worse that can happen?

All Your Browing History Are Belong To Us.

Well, apart from the privacy invasion (your search history can be determined by the urls in your browser history), there are all sorts of nasty things that can go on.

One example is the fact that some sites are designed with appalling security - explicit storage of usernames and passwords in the url is not unknown. Nasty. If you happen to use the same username and password across multiple sites, then the malicious can try out the unencrypted usernames and passwords on other more secure sites (that listed in your browser history) potentially yielding some very bountiful results. Nasty, nasty.

In Attention Data We Trust

The attention landscape is and will be full of privacy concerns - there is an ongoing balancing act between potential abuse of the data 'submitted' and the potential benefits in providing that data.

At the end of the day, it is down to individual risk assessment. As a 'customer' you need asses whether you can trust the site to provide you with more relevant experience based on your attention data and do so in a safe manner.

The browser history case is the one of the more poignant examples in this regard. We're not even talking about malicious sites here. You not only need to trust the site owners in terms of proper use of the data that is collected (their privacy policy and its adherence) but also trust their staff's competency in securing their infrastructure - if the data store that holds your data is not properly secured with proper physical security, rigorous policy and processes then you risk saying bye bye to your data (a la Mastercard).

My Attention writings

Tags: attention security trust mydata

Recommendation and the Harry Potter Problem

alexbarn — Thu, 23 Mar 2006 19:25:00 GMT

Greg Linden used to work at Amazon.

From time to time he'll recount one of his 'early days at Amazon' stories on his blog giving some insight into the development challenges in building the world's largest online shop.

In his latest post, Greg talks about the "Customers who bought this also bought" feature, internally known at Amazon at the time as the 'similarities' feature. One of the early problems with 'similarities' was the 'Harry Potter' problem.

See, everyone and their dog bought Harry Potter:

"This kind of similarity is not very useful. If I'm looking at the book "The Psychology of Computer Programming", telling me that customers are also interested in Harry Potter is not helpful. Recommending "Peopleware" and "The Mythical Man Month", that is pretty helpful."

So he was asked to fix the Harry Potter problem, which he did.

The fix not only provided a better shopping experience at Amazon, but also resulted in Jeff Bezos walking into Greg's office, bowing on his knees and chanting 'we're not worthy, we're not worthy'. How cool is that?

One interesting point to note here, maybe even a self-referential observation. And that is I found Greg's post about recommendation systems via the personalized version of Findory which he developed, itself a recommendation system that does a good job of avoiding the Harry Potter problem.

10 Random Thoughts. # 2 Repetition engines

alexbarn — Fri, 17 Mar 2006 18:50:00 GMT

10 Random Thoughts.

# 2 Repetition engines

New media is rapidly becoming like old media.

'Memetrackers' (newstrackers) are dragging 'new media' into the land of the repetitive, the shallow and the incestuous.

They are not what I meant when I described attention engines. I want my niches, not personalized mass new media.