A year ago I pointed to a problem regarding the fragmentation of the feedosphere, the scattering of our digital streams across numerous social media - Flickr, del.icio.us, blogs, etc. I was hoping to compile an organized list of feeds so that you could easily scan the published online activities for participating Irish bloggers.
Lately people like Jeremy Keith, Edward O'Connor, Chris Messina and Sam Sethi have been addressing this issue with something they call lifestreaming, which pulls one's scattered feeds back into a single spliced digest. Very nice, except it's a band-aid solution to an aging architecture - the app-centric web. Which evolved from the page-centric web.
But the Live Web , the social web, is people-centric and a new generation of aggregator/annotators (e.g. Google ReWriter) will generate lifestreams at source. Or more precisely they'll generate uniFeeds at source to which our web apps will subscribe and render the relevant parts. Why browse the web when the web can browse you? Lifestreaming is a valiant effort to rein in our scattered digital fragments but OpenID associated uniFeeds will be a single source of all our digital outputs - the ticker tape of our lives.
Technorati Tags: lifestreaming, unifeeds
March 05, 2007
Surely, one thing that could be useful would be if both HTML, RSS and OPML had a way of adding data to point people to "my other feeds". That way, when you are browsing a web site and you click subscribe, it could ask you "do you just want to subscribe to Eirepreneur, or do you want to subscribe to James' Flickr photos and bookmarks?"
This is where microformats and the SemWeb comes in. XFN already has the "me" class - so that you can point people to pages that are mine. FOAF has a property called OnlineAccount - which lets you specify other online accounts as part of your profile. I could see, for instance, an OpenID provider which pumped out a FOAF file listing all your different online accounts - and that would be trivially easy to turn in to OPML.
In fact, here's another idea. Imagine if you could subscribe to a feed of feeds. It would be an RSS file which could contain enclosure links to other RSS feeds. You would tell your newsreader that if James includes a feed in his feed to automatically subscribe to it. Why not just use OPML? It seems that OPML adoption, beyond primitive import/export, hasn't really been taking off too well. The idea of feed-in-feed would be that if you launch a new blog, you could announce it on your current blog and your visitors would be able to be automatically subscribed - if that's indeed what they instructed their newsreader to do.
Posted by: Tom Morris | March 05, 2007 at 03:19 PM
Very interesting ideas Tom, I certainly hadn't thought about things in that way before. However, I still feel that there's something fundamentally wrong with scattering our feed fragments all over the web regardless of clever ways to announce, connect and subscribe to them.
It's like telling my friends that I'm going to clone myself and if they want engage with 'funny' me they can chat with my 'funny clone', if they want to listen to pensive me they can subscribe to my 'pensive clone, etc. Obviously that's not the way the world works - anyone who wants to tune into my stream of consciousness just listens to (the one and only) me! They can choose to let some things in one ear and out the other if they're not interested (as we all do from time to time) and then 'render' the stuff that is relevant to them.
Sorry if the analogy is a bit of a strech (;-) but the point is made to reinforce my view of the changing architecture of the web - that the web revolves around the individual.
Posted by: James Corbett | March 05, 2007 at 03:38 PM
A few thoughts. FOAF/ RDF is not happening on internet. RDF may be doing something somewhere on some back office server. They have been talking for years of embedding RDF in HTML. It is still working draft. I don't see it "ever" hitting the streets. I took my FOAF file down 6 months ago. It never got a hit. XFN/ microformats is a far more consistent anyway IMHO as it embeds that type of metadata in HTML. Microformats/ XOXO is similar to OPML and can be embedded into HTML. But hardly anyone uses it. I think Lifestreams are cool ... but google rewriter is a great 'idea', but is anything like that being created anywhere? (FOAF/ RDF makes me froth in the mouth, The rest are sincere non-feisty questions.)
Posted by: Don | March 05, 2007 at 08:36 PM
Sorry to turn this in to an RDF fight, but RDF was finalised as a W3C standard in '99, but it took a lot longer for it to become adopted as general practice by web designers. Compelling use cases sold it, not it being a standard. I do not see why SemWeb technology won't take a similar adoption path - the technology for embedding and extracting semantics from HTML is reaching a point where it is usable.
That's outside the issue though. I'm just writing a script at the moment that pulls in a web page, extracts the feeds, follows the links marked with XFN and pulls the RSS feeds from there, as well as checks for FOAF files and looks up OnlineAccounts in those, compiles the lot in to an OPML file and sends it off to FeedBlendr. What does this mean in practical terms? You find a blog - you click a button in your toolbar, it checks the page through for both microformats and other semantic elements, hunts down the feeds and you've got a 'lifestream' there as both an RSS/Atom feed, and an OPML file which you can share with others or use in readers which support subscribing to reading lists.
Whether the current SemWeb technology gets us there is irrelevant - computers that can do things based on the type of relationship between one object and another will make browsing subscription content much more interesting. James tells me that he wants the SemWeb and the SynWeb/LiveWeb to talk to one another. This is a primitive way that they could.
Posted by: Tom Morris | March 06, 2007 at 09:22 AM
Whoops. What I meant to say was that *CSS* became a W3C standard in '96. It took a long time to get to mass adoption. If you had said in 2000, "Wow, CSS has been out for so long, it'll never get mass adoption", you'd be wrong. I'm not sure why SemWeb technology is any different, except that it's significantly more complex and thus the timeline of adoption should be extended further.
It looks like this year, more and more steps towards adoption are going to be made, but we've still got a long way to go.
Posted by: Tom Morris | March 06, 2007 at 10:08 AM
Tom, I don't want to do any mud slinging. I am frustrated at how long some of this stuff takes to hit the streets. But also when it hits the street it is not always easy to use. Do a google search on "SOAP sucks" or "RDF sucks", it is full of bloodied stories. Anyway web 2.0 is taking the same ideas and making them work. Thats what counts.
Posted by: Don | March 06, 2007 at 08:32 PM
Interesting debate lads - I'll watch from the sidelines if you don't mind ;-)
Posted by: James Corbett | March 08, 2007 at 09:49 AM
Videntity.org is already doing the OpenID to XFN to FOAF thing... and then you have various other attempts, like Jaiku and Profilactic auto-generating these lifestreams.
I actually think that the solution is more in something like managing your attention data as you do your bank account -- and frankly, I have a few accounts at different banks for different reasons -- and that's how I like it. A uniFeed seems redundant -- let apps do what they do best and then build various value layers on top of that.
Or, to put it another way, attention is currency; I should be able to dictate how I save and spend it.
Posted by: Chris Messina | March 10, 2007 at 01:08 AM
Which format would you use for the Attention Data side of things Chris :)
Posted by: Chris Saad | March 10, 2007 at 01:59 AM