Mashable OPML attention

Pete Cashmore has uploaded his OPML file / blogroll into too.

Worth browsing /  importing and seeing what his RSS subscriptions is you’re interested in Web 2.0 / API / Mash ups. He’s also been following the OPML / Attention conversation:

“For those who haven’t grokked it yet, OPML is essentially a way to exchange lists of RSS feeds between RSS aggregators. (Alex explains it here). However, it has the potential to do much more – since your RSS feeds are indicative of your interests, your OPML file is a major source of attention data. “

Michael Parekh picks up on Raymond’s post I pointed to yesterday and asks:

“While I agree with how OPML files can become as important as other real-world documents in mainstream life, I do have a pedestrian question.

The discussion above assumes there is one central OPML file that aggregates all this user attention data over time.

Reality though is much messier.

As I look at my own situation, I have “OPML assets” scattered all over the place.

The biggest one is at Bloglines as I described earlier.

But I’ve also got a burgeoning OPML file over at Yahoo!’s “MyWeb”, along with other increasingly important ones over at

And as other tagging and RSS aggregation sites proliferate, I’m likely to have another half a dozen more at places like Digg and others.

Where is the centralization?”

Centralization of all these OPML files and RSS subscriptions? Answer: an OPML file. An OPML file can point to other OPML files and urls (that each in turn can include RSS urls and normal links). So you could have your ‘master’ OPML file, that points to all the others you’re maintaining on other systems / apps. The interface for all this isn’t there yet. comes close but I don’t think you can include / point to other OPML files (Bernard? and Pieter?). If you could that would be very useful. I managed to get OPML Editor installed now: that’s another way of managing the centralization, so more OPML playing ahead for me.

But not yet. Time for turkey now.

Tags: Web 2.0, OPML, Attention

Comments (6)

  1. does support inclusion! Just add a link that points to a opml file. The inclusion rule for now is that the file you point to ends with .opml => type="link" url=""

    Normal links are visualizez with the orange dot, links to opmlfiles with an arrow (=>)

    Dave Winer has proposed a new approach to inclusions with an extra type for inclusions => type="include". I’m in favour of type="opml". This will make it possible to include opml-files that do not have the extension .opml. Which makes it easier for webdevelopers to generate opml from a script => opml.php or opml.aspx…

  2. btw, its Pieter (pronounce like Peter) 😉

  3. MSDNArchive says:

    thanks Pieter! (sorry about the mispelling!)

  4. I hope Alex doesn’t mind if I ask here Pieter but will you be adding AJAXy features such as drag and drop rearrangement of directory entries as you can do in OPML Editor? That would make OPMLmanager even better than it already is 🙂

  5. centralisation is not important. its replication thats important. thats how come i can’t belive you didnt point to SSE here…

    its a quadrangulation of SSE, attention, RSS and OPML that creates the opportunity, no? or that is a triangulation of specs leading to attention?

  6. MSDNArchive says:

    Ok James, I see your point.

    Let me explain.

    In this context, ‘centralization’ is meant in a virtual sense. The nice thing about OPML is that it can be (but doesn’t *have* to be) a distributed model: the first OPML file sits at url (a) and can point to another OPML sat at url (b), so as the url at (b) is updated the one at (a) is too (at least at the point in time that (a) is pinged). So SSE in the scenario I descrbied is not required: thing ‘sync’ by design. SSE solves other problems.

    SSE can come into play as you imply when other data sources need to write to the OPML or RSS. You’re right to point this out.