The elmcity project’s single biggest hurdle continues to be a conceptual one. People mostly lack the intuition that it’s possible — never mind easy and free — to publish data that can syndicate. In response to an earlier item on this topic, Stefano Mazzocchi (Cocoon, Simile, and Google Refine) offered some thoughts which I’m sharing with his permission:
A few weeks ago, my in-laws were visiting. She is a pretty famous book author and we were talking about how technology could bring value to her workflow (she is not flat out an IT luddite but close enough).
She just created her first web site and asked me how she could promote it on search engines (classic newbie SEO question). She was not interested in the mechanics at all, she just wanted more exposure.
We talked about Twitter and about how publishers and authors use it to promote themselves and engage their audiences. She thought all this was very “Hollywood” and not her style at all. But I showed her that you don’t need to use Twitter that way, you can just mine it for your ego network. Then I explained how I set up all sorts of traps around the web, with newsfeeds, and how I use Google Reader to aggregate them all for me.
I showed her right there and then. Searched for her new book name on Twitter, clicked on the RSS feed, did the same on Google blog search, Google news search, and voila, her personal PR aggregation network was born.
She was completely blown away. She didn’t know any of this was even remotely possible, yet, once explained, it make perfect sense. It’s like having personal agents watching everything that goes on and sending you the information. Email versus RSS doesn’t make any difference to her. As long as she has a place to go and check out what others say about her, she’s happy.
My take: no tech-unsavvy person thinks it reasonable to have a personal agent that does, for individuals and for free, what gigantic organizations struggle to do every day.
The fact that Google can search 15 billion pages in milliseconds doesn’t faze them as much. If librarians can do it, so does Google. Big deal.
But personal agents constantly working in the cloud for you? It doesn’t even show up in the realm of possibilities.
If Stefano’s in-law were on Facebook, of course, she’d be getting a sense of what it’s like to have one of those agents in the cloud. Her activity stream would magically be visible to friends, and their reactions to it would magically be visible to her. That’s why I often say, nowadays, that Facebook is a great set of training wheels for the pub/sub network.
But Facebook isn’t, yet, a place where people can learn how to publish data that syndicates beyond Facebook. It’s possible, as I discussed in Heds, deks, and ledes, to post public events on Facebook in a way that can be discovered by an elmcity hub or by some other agent. If you don’t know such agents can and do exist, though, you’ll never stop to think about whether they’re actually finding your data — and if not, how to make sure that they do.
One of the key points embedded in Stefano’s parable is that his in-law didn’t have to do anything special in order to be able to find the web’s reaction to her book. To the extent that her name and the name of her book are out there and indexed, they provide good-enough hooks for search aggregation.
Over time, of course, the efficacy of these searches will decay. I’ve watched this happen with my own name. Years back, my stuff was pretty much the only stuff that a search for Udell would find. (There was even a time when the first Jon on Google was me, not Jon Stewart!) Then my wife began showing up, along with a whole bunch of other Udells. So I tuned my filters to Jon Udell, and they work better for now, but there are other Jon Udells and it’s only a matter of time before that namespace gets cluttered too.
In order to reliably find stuff about me, I need filters tuned to aspects of my identity: my domain name, my Twitter handle. Eventually Stefano’s in-law may reach the same conclusion. She may realize that posting to her website is more than a way to share her thoughts with the world. It also enables the world to react to her posts in ways she can, in turn, discover. At that point she may start to see why it’s important to actively colonize parts of the web that are, or can be, bound to aspects of her identity.
The elmcity project, similarly, invites promoters of public events — and communities at large — to colonize the web in ways bound to those individual or group identities. When you produce a calendar feed that flows through an elmcity hub, you’re not just helping to populate that hub. That feed is attached to your own site and, in theory, is directly discoverable there. In practice, though, there are aren’t yet good methods of discovery. We don’t yet have, for iCalendar, an autodiscovery mechanism like the one we have for RSS. That’d be easy enough, as Mark McClaren suggests:
Love it or hate it, iCalendar is the pervasive calendaring format. If we can enable RSS autodiscovery then why not do the same with iCalendar feeds. Adding one line of code would make it easier for people/machines to subscribe to an iCalendar feed.<link rel="alternate" type="text/calendar" title="iCalendar feed for example.com" href="calendar.ics" />
It would also be really helpful to be able to bind locations to events in a discoverable way. To that end I’ve recently enhanced the HTML rendering of elmcity hubs. Now they include what Google calls rich snippets, using the RDFa-style markup documented here. The snippets include latitude and longitude coordinates derived in one of two ways:
1. Per-event. There are several ways that an event can show up bearing latitude/longitude values. The vast majority of such events will be those coming from Eventful and Upcoming, both of which services provide lat/lon values via their APIs. There’s also a GEO property defined for iCalendar, and some iCalendar producers use it to geocode events.
2. Per-hub. Although most iCalendar producers don’t use the per-event GEO property, elmcity hubs know their own locations. So events that lack specific lat/lon coordinates inherit the locations of their hubs.
It’s going to be a while yet until folks like Stefano’s book-writing in-law start to realize they can, as Kingsley Idehen nicely puts it, master their own seach indexes. But sooner or later they’ll realize that it’s possible. Likewise, it’ll be a while yet until promoters of public events realize that the event data they push to their websites can not only feed pub/sub networks, but can also feed location-aware search engines. I’m a patient man, though, and I do expect the seeds I’m planting to grow and eventually bear fruit.