So. Having done some research: a number of sites offer calendar hosting and search. Well and good. But obviously not enough to get people very excited: these sites are mostly fairly low-key, even though they offer some fairly high-value data (calendars for sports rosters, national holidays and so on), that in a world truly turned-on to the power of calendar pub/sub, you would reasonably expect to find not tucked away on a hosting site, but regularly updated on the websites of the related organisations, teams or governments. There’s a real gap between potential and perceived value.
What’s missing? Or a better question: what’s needed to push this into broader consciousness? Two inter-related aspects: passion and personality. People care about calendars because they care about stuff which happens: bands go on tour, teams compete in championships. People will pick up on technology to the extent that the people who are doing the stuff that people care about (let’s generalise them as actors) pick up on it. we don’t need more sites hosting random collections of calendars, we need the people who do stuff hosting their own, so the people who are passionate about what they are doing (lets call them the fans) can keep closer in touch, more easily.
But it’s rare than fandom is limited to a single actor — most fans care about related clusters of actors and what they’re up to. Currently fans who want to track a number of actors’ calendars use client-side aggregation/layering in desktop software like iCal, giving them their own personally aggregated ‘calendar-of-me’, reflecting their varied passions and predilictions.
And there it currently stops.
As if we had never had IMDB, playlists or Amazon lists, or any of the other network-based, edge-fed collaborative filtration systems. What’s missing: internet-based publication of those personally-aggregated clusters of actors and events back online, so we can all benefit from the incremental value added at the edge by people who care about stuff, and aggregated back into the network (the purple arrow in the above diagram). With that feedback in place, we could start building systems that harvest value from network characteristics and patterns: popularity, clustering, geospecificity. The (small) benefit that each edge user accrues from their own desktop aggregation would be no longer be lost out at the edge, but instead become part of (yet another) network of live networks (like all the others with which we are now familiar: CDDB (as was), page ranking, LiveFM, mp3 playlist search etc.). And I think aggregation does add value for calendar systems. Amongst other network effects, it’s easy to see how preferrential aggregation would quickly give rise to ‘instant fame’ for fans whose calendars were really useful to others: imagine city-wide gig guides for specifc genres of music. Or the stickiness of such guides if sponsored by famous DJs or music critics. There’s something to this. More thinking to do
conceptually, you publish your events to ICdb (Internet Calendar database al la IMdb — though perhaps more of a virtual, distributed-store aggregator vs. IMdb centralized model) and that data is indexed by submiter (fan), calendar name, event types etc etc … you then set your own calendar to subscribe to ICdb based on various filtering criteria (rather than just subscribing to certian calendars). so lowpass filters needed to avoid feedback loops i guess.
Comment by Tim — Saturday 18 December 2004 @ 9:43 pm
yep — i’m thinking even a few more ways to slice it — raw calendar data relates to events — events can be not just POSTED but REVIEWED. So you can ,as a fan, get fame through being a good reviewer as well as a good REPORTER — the number of events to REPORT is finite, but the number of OPINIONS is infinite…
SO you could subscribe, say, to the ‘london new music channel, but only events performed by actors who have previously got a 5 star rating from a particular reviewer…
The three things i think missing from hte current crop of cal hosts are:
- understanding of the need to create a MARKET for this stuff: not a tehcnical challenge, its a business one and involves a brand and lots of PR
- understanding of the importance of collaborative filtering (all the stuff above)
- understanding of the importance of server-side aggregation and re-publication as part of thecollaborative filtering.
Comment by darrell — Thursday 23 December 2004 @ 9:28 am
indeed.
the cross-connections of calendars to other data sources has interesting collaborative aspects. eg lets say you have a film calendar where (unlike live performace) the “events” have probably already been reviewed by others in IMdb — you can get instant calendars for currently playing films that match your taste/needs.
or book channels with reviews databases that then could match live events (books released/authors speaking etc)
and then of course reviews of (live) events feed-back into the “static” reviews of authors/creators/films/books etc
feels like a hyperlinked event listing wherein the links are polled to generate a metric for YOU. bayesian filters.
Comment by Tim — Thursday 23 December 2004 @ 4:21 pm
yep — that’s a pretty essential difference between a calendar-of-me service and pure hosting… and easy to see how you extract value both inside the system (directly for users) and outside it (for sponsotrs/advertisers etc)
Comment by darrell — Thursday 23 December 2004 @ 5:17 pm
and also — the CDDB/IMDB model has the extra added coolness of appealing to fans who are passionate RIGHT ACROSS the bellcurve — which means you get a much wider range of events fed in than the listings magazines etc would ever see… in fact i suspect that going for the geeks, freaks and fringers first is a good strategy — getting the stuff in the centre of the bellcurve is easy, but a lot of the value will come from the collaborative filtering between the fringe and the ‘hump’
Comment by darrell — Thursday 23 December 2004 @ 5:23 pm
and i guess live feeds from stuff already happening, should anyone be into that… quite like the idea of the calendar day squares serving up thumbnails of video content… later later later
Comment by darrell — Thursday 23 December 2004 @ 5:36 pm
Have you seen the rather lovely table about the relative sizes of various kinds of searchable data sources mentioned on Technology Review? Interesting in that it doesn’t include event data at all…
Comment by darrell — Sunday 26 December 2004 @ 9:25 pm