the onda

Doc has a characteristically thoughtful piece over at Linux Journal on the need for web service creators to be good members of the ecosystem and support APIs that allow data to be federated across services. He uses the example of his Flickr stream being usable at Tabblo (which is a testament to how good and ground-breaking Flickr has been in this regard), making the point that it benefits him as a user when all of us "vendors" play nice and respect both his data— and more importantly, his time. It is just plain silly to make him waste time with the asymmetry in broadband today, having to upload all of his stuff again.

I've written before on how business models that are at odds with this fundamental respect for the user are likely to be doomed (see all of the stuff on the roach motel), so here I'm going to take another angle. Data acquisition, whether it is importing contacts in a social network or uploading assets in a photo site, is just not something that we need to keep on re-inventing over and over again. Let me stick with the latter example for a moment: outside of the Flickr API, at Tabblo we have 7 other methods for ingesting digital photos. Each of these requires constant maintenance as its client environment/runtime changes, something that I think is just too much of a support problem for any small team to take on. And in fact, these days each of our rich uploaders is in some less-than-optimal state: from "outright busted" to "works most of the time."

Contrast this to Flickr which managed to solve the uploader maintenance problem by crowdsourcing it and letting passionate community members maintain all sorts of different clients with varying degrees of success. Not everyone can be a Flickr, but we should all be able to leverage their success, and instead spend out own cycles thinking about how to build the next layer of value.

That said, what I'm having trouble reconciling is how to build simple user experiences in this new layered and federated world. Imagine if we asked the average mainstream user who comes to Tabblo to go and register at Flickr, upload their photos there, authenticate Tabblo as a trusted service, only to then be able to get into their story-telling process. Nightmare, plain and simple.

One potential solution: I'd like to see a white-label services that could be wrapped by webapp builders for core pieces of functionality. To continue the upload example: why doesn't Amazon, or some enterprising entrepreneur looking to build on the cloud computing infrastructure at Amazon, build out a full suite of well-supported file uploaders, along with an associated S3-backed storage infrastructure for everything from photos to videos. By focusing on just the upload experience, this effort could just nail it for all the rest of us— building plug-ins for our favorite apps, clients for our favorite platforms, and even specialized hardware for events and community activities. In Doc's VRM world, such a company might even be able to charge the enduser a nominal fee for pipe and storage, so long as its service integrated easily with enough of the interesting webapps.

You listening lazy web?