OPML and the road to Attention data: Progress

Steve Gillmor informs us of Dave Winer’s OPML namepace news:

“Basically, I think OPML should have a way to include elements from other namespaces, much as RSS 2.0 does. Developers should, whenever possible, use capabilities that are already in use by others, or included in the core spec, or recommendations or guidelines. I think RSS 2.0 would have done even better if this had been spelled out in the spec, so I plan to include this advice in the new OPML spec.”

This is good news. All of a sudden, the idea that OPML can become a format for attention data is a step closer to reality. I’ve been pushing for this so am pleased with Dave’s news. (Nick Bradbury and Steve Gillmore have both picked up on my earlier call to restart the conversation around Attention.xml and OPML). One of the sticking points to making this happen was the lack of support for namespaces in OPML.

In reponse to Dave’s OPML namespace update tonight, Nick Bradbury proposes we keep things simple to start off with:

“Although there’s a lot of attention data that could be stored in OPML, my recommendation is that we keep it simple – otherwise, we risk seeing each aggregator support a different subset of attention data. So rather than come up with a huge list of attributes, I’ll start by recommending a single piece of attention data: rank.”

Nick then invites us to think: “Beyond rank, what other attention data do you think aggregators should collect? And how should they use that data to serve you better?”

Beyond ‘rank‘? I have some thoughts on this.

The Vote Attribute

Some have been pushing for ‘clicks‘ as an attention data attribute. This clearly has value, but is complex for the resons I give below. There is a simpler way that I think can bring quicker results to market.

I think what could bring more immediate value to us is something along the lines of a ‘vote‘. It is a word Robert Scoble uses when he talks in the context of blogging. By linking to a post, he his providing a vote for a post. He’s adding ‘search engine juice’. This is very different to a click. The problem with using clicks as a proxy to attention is that a click doesn’t necessarily mean I’m interested. In fact there are many problems with using clicks as an attention data bit (they keep cropping up in conversations I’ve had). Simple example: It might be that I clicked, and realised the post was not for me.  Should this count? The ‘solution’ to this specific ‘click’ problem is to measure the amount of time a user spends on a page, and so on and so forth. This click route gets quickly convoluted and as you might see this is anything but a clear path to understanding whether a bit of content is actually of value to the user. And how do you quantify it…what about privacy?) It is do-able and this is what drives recommendation engines at places like Amazon, but at this stage of the Attention game we should gravitate toward the simple where at all possible. Quick wins and all that.

My view is that a simpler, neater, easier and instantly meaningful attribute to implement is a ‘vote’ type attribute. By vote I mean I like this item, page, post, or podcast, or whatever, so I vote for it. By making a deliberate action in casting a vote against a piece of content, I am positively registering my interest (we could have negative one too, but that’s another kettle of fishy things). No fuzziness involved. A vote has explicitly more value than a click. You can think of this being similar to bookmarking (traditional or social along the lines of Del.icio.us).

How do we use this?

So how would this vote attribute be used? In the case of an RSS / feedreader, the votes can be collated, stored and then shared within my OPML attention file. This is what Nick proposing with rank of feeds.  You could do this with on local desktop readers such as Nick’s FeedDemon, or with webbased readers, such as Bloglines and Windows Live. Nick has more on how this would work, but the idea with rank is that it would be implicit data collected via the use of the application. And this works well at the feed level.

A vote would work at the item level. (I repeat: by item I mean RSS item, webpage, blog post, podcast, or video or whatever – if it has an url it can be voted for). Voting would be explicit, requiring a user action, maybe a quick check of a box. This vote data along with feed rank data would be probably enough data to do some very interesting things. The nice thing about all this attention.xml / attention / OPML discussion is that application developers can compete around the applications’ algorithms and other data that turn this attention data into useful things –  Attention engines, book and music stores could all allow the import of my attention (OPML file) or point to my OPML attention url to allow more relevant experiences.

Anyway, these are just my initial thoughts given the OPML namepaces news. We’re really moving now…All very exciting!  Thanks to Dave, again.

Update: Nov 18 2005:

Tags: Web 2.0, OPML, Attention

Comments (12)

  1. Are you taking the piss alex? didn’t i correct you on gilmoor before? check yourself- two different spellings in two pars. no wonder you’re so happy with sloppy specs…



  2. oh yeah- the voting stuff. are you aware of the lessig vote tag idea from the last election?

    it set off this take on attention before i ever heard of it


    please read it, if you didn’t the first time, and let me know what you think

  3. MSDNArchive says:

    LOL! OK, I’m leaving it there for posterity.

    Will check out your post, thanks.

  4. Stewart says:

    Does Rank suggest a historical rank, number of clicks over time, and if so aren’t you effectively voting by visiting the same resource again and again?

  5. Isn’t including the content (RSS feed, whatever) in the outline element a "vote" for it in the first place? You’re effectively voting for that content, and not voting for content that you don’t put in. What meaning does it have to not vote for something you include? Then you have three states: not included, included and not voted for, and included and voted for. All you’re doing is increasing the number of attention states from two to three. If that’s what you want, then fine. I could see people voting for everything though, so you’re left with effectively two states again and it’s pointless.

  6. I’ll echo Paul’s comment by saying that subscribing to a feed is automatically a vote for it. It might be worthwhile for aggregators to also keep track of feeds you’ve unsubscribed from and flag them with "vote-against" – this way the importing aggregator would know not to recommend feeds you’ve already seen and removed from another aggregator.

    BTW, Attention.xml includes a rev/votelink attribute:


    Is this more or less what you’re proposing?

  7. MSDNArchive says:

    Paul, Nick. You talking at the feed level. As I tried to stress in my post, I’m talking at the item level.

  8. Voting at the item level seems useful, but given how many items we read each day, I’m not convinced enough people would take the time to manually vote for individual items. However, if the action wasn’t an explicit "vote" it might work – for example, if you flag an item in FeedDemon, it could be considered a vote for that item.

  9. MSDNArchive says:

    Yes, Nick, that would do the trick!!

  10. Glad you agree 🙂 I think in general, most people don’t want to take the time to rank items, so tools such as FeedDemon need to infer ranking (or voting) based on the user’s actions.

    That said, explicit voting at the feed level might be practical, since people may be more inclined to say "I like this feed" than they would "I like this post."