Bomb the cluster, plant memetrees instead
The launch of Personal Megite, featuring "memetracker" reading lists seeded from OPML files provided by Robert Scoble, Doc Searls and Richard McManus, has caused a bit of to and fro, not the least with Richard getting criticised for praising the service. Nik Cubrilovic, now of storage startup OmniDrive but fresh off trying to build a memetracker of his own called Pleech, beseeched those in the increasingly crowded space to forget personalisation. Gabe Rivera of Memeorandum chimed in supporting Nik on Scoble's blog.
In Matthew Chen's defence, he has actually released something. That's more than most have done lately (mea culpa!). I have watched Megite lately, and it has gone through some interesting mutations. Its current incarnation looks almost exactly like Memeorandum, all the way down to the boxes with plus signs in them to expand secondary blog lists, as I predicted. I suspect, nay hope, that that is merely one step on the way to something newer and better.
IMO, personalised clustering is not "too hard". The technology is present to be able to do it, even if only as a mashup - of course, if you're Technorati-level or higher you can mash your own DBs up. The problem is in finding an application that users want, to make the phrase "personalised memetracker" mean something - which might be different to what Megite is doing and what Gabe and Nik have tried to do.
What might be a more compelling application is to seed the memetracker with only one source: the user's own blog. Who among you reading this keeps track on Google Blog Search and/or Technorati of links to their blog? Instead of a Mememorandum-style clustered listing of primary stories with multiple secondary stories below it, the data could be displayed in what might be called a "branch-and-root" or "memetree" structure, with each of the user's blog entries as the base, outgoing links from that entry as the roots, and incoming links as branches. It could be implemented as simply as adding a second set of links above each base snippet in a Memeorandum template, or it could get all VMLish with multidimensional link trees sprouting up like creeping vines.
If done right, I think it could capture the considerable market for ego feeds, which is currently being serviced in a slipshod, scattershot manner. Technorati and countless other present and future aggregators have all of the plumbing behind this concept covered. All it needs is one prototype.
In Matthew Chen's defence, he has actually released something. That's more than most have done lately (mea culpa!). I have watched Megite lately, and it has gone through some interesting mutations. Its current incarnation looks almost exactly like Memeorandum, all the way down to the boxes with plus signs in them to expand secondary blog lists, as I predicted. I suspect, nay hope, that that is merely one step on the way to something newer and better.
IMO, personalised clustering is not "too hard". The technology is present to be able to do it, even if only as a mashup - of course, if you're Technorati-level or higher you can mash your own DBs up. The problem is in finding an application that users want, to make the phrase "personalised memetracker" mean something - which might be different to what Megite is doing and what Gabe and Nik have tried to do.
What might be a more compelling application is to seed the memetracker with only one source: the user's own blog. Who among you reading this keeps track on Google Blog Search and/or Technorati of links to their blog? Instead of a Mememorandum-style clustered listing of primary stories with multiple secondary stories below it, the data could be displayed in what might be called a "branch-and-root" or "memetree" structure, with each of the user's blog entries as the base, outgoing links from that entry as the roots, and incoming links as branches. It could be implemented as simply as adding a second set of links above each base snippet in a Memeorandum template, or it could get all VMLish with multidimensional link trees sprouting up like creeping vines.
If done right, I think it could capture the considerable market for ego feeds, which is currently being serviced in a slipshod, scattershot manner. Technorati and countless other present and future aggregators have all of the plumbing behind this concept covered. All it needs is one prototype.
posted by Paul Montgomery | 10:24 pm
0 Comments:
Post a Comment
<< Home