The Future of Web APIs

We had an interesting day at the Future of Web Apps “Summit” yesterday - some very interesting speakers, some less engaging presentations, and some interesting ideas to take away and mull over.

I found it particularly interesting that a good number of the speakers were pushing the idea of collaborative web service APIs - by that I mean people like Flickr releasing an API and other people building new and interesting sites and services off the back of that window on the data.

For me, this raised some interesting questions for any business looking to use external data in this way, some of which we mulled over whilst enjoying a leisurely lunch in Wagamama after bumping into a nice bloke called Jan from some little foreign browser company… I think it was called “Opera”…

Anyway, enough name dropping…

The two main issues that we came up with were:

1. What happens if the top-level API provider decides to close or monetise their API?
2. What happens if the top-level API provider decides to add advertising to the data before it reaches you?

For any business that uses third-party data, there is a definite problem with the evolving “fast and loose” model of building on public APIs - resilience.

With many existing case of the use of third-party data, there are contractual fall-backs in place - SLAs - that define what level of service can be expected from each party.

With the current drive to build on top of public APIs, there is seemingly less focus on this protection of the investment involved, and several speakers on the panel discussion weren’t convinced that they were necessary. I’d say that that view was a little irresponsible.

What happens if Flickr or Google Maps decide to close access to their APIs?

Their services are consistently presented as “Beta” (another contentious issue, especially if you happen to be talking to David Heinemeier Hansson of 37 Signals) and the fiscal giants that are Yahoo! and Google could readily close off the API for a full software release without even a “thanks for the help” to the developers who have found new and interesting uses for their data and publicised the service in the process. And any company that uses that data as a core part of their offering is up a certain well known creek, sans paddle.

And what if Google decides to charge for access to their mapping doo-hickeys? Time to re-draw your cash flow projections, people.

And what if Google decide to leverage the synergies (I had to get a crappy business buzzword in there somewhere) of their position as a targetted advertising company and their ground-level mapping data from Google Maps to offer businesses the opportunity to have a funky little icon show up on their location whenever someone zooms to street level on their block?

Big deal, you may say. But what if you’ve used the GMaps API to stick handy maps onto a public sector or government website? What if they are fundamentally opposed (or even legally precluded) from carrying adverts? Either the maps go bye-bye, or you stump up some dollars to get a “premium” API service without the location-based advertising - more cash-flow re-forcasting.

Now don’t get me wrong - I’m not against the use of public APIs, and I’m not saying that either Yahoo! or Google will definitely go down any of the routes mentioned, but it’s very easy and very logical to suppose that they might.

For the record, I think that the opening up of APIs is an excellent idea. I fully support the move away from compartmentalised, ruthlessly guarded and viciously priced proprietary data towards a service-based model which promotes what is possible, rather than what is permitted.

• Thursday 9th February
• Posted by Andy Hawkes