Google maps, Flickr, Ebay, and other web services with APIs are pulling the relevant platform away from the desktop and toward the web.
Still, the network effect of powerful, privately owned web APIs is potentially as dangerous as the network effect of Microsoft’s desktop APIs. On any given day, Google or Ebay have the right to change their APIs and make life difficult for their developers. They have the right to change the terms of service, and increase prices on services that their developers depend on completely.
The lockout effect could be even worse, because Google and EBay own the servers, and changes can take effect in real time. When Microsoft bakes DRM into every copy of Windows, users don’t need to upgrade their PC immediately. But if Google or Ebay changed terms of service, those dependent on the service would need to comply immediately.
Google and Amazon, and Ebay’s big servers are a big deal. A web service can start small. But once service becomes popular, it takes a good amount of capital to complete. Currently, competition between GOOHOO and AMABAY are keeping things lively. But oligopoly could lead to complacency and extractive economics, as in other industries.
The owner of a dominant API/service is in a very powerful position. Google has the ability to adhere to its corporate slogan, “do no evil.” That ethical stance does make a real difference. A powerful ruler can choose to be a benevolent dictator or a tyrant. But the temptation is there for power to corrupt.
I can imagine a way out of this oligopoly bind.
What if there was peer to peer for web service requests. Many small servers could run the popular service, and publish their availability. When a client issues a request, the request would be taken by an available server. This wouldn’t work for services that require a pre-existing content store (like maps?). But it would work for services that require large amounts of individual content (like calendars?).
Maybe the technology already exists somewhere, and is waiting for the killer app. Maybe I’m missing something — this is just musing outloud. What do you think?
How about integrating with redundant web services. For example, if you really on paypal for payment procesing, hook into a merchant account.
I think the key to all of this will be that there are choices between web services. That, if we want to build an application off of maps, that there are two services that are interchange-able.
Right now, there isn’t much competition in the web services business. But, I expect that it’ll be coming shortly.
Wow. I really should reread my comments before I post.
Right now, I don’t see the market power that would drive web services APIs to standardize.
There are a few powerful providers of major APIs, and many small developers. Standardization benefits developers but not providers. Until there are users or user groups that are as powerful as the API providers, I don’t see standardization happening.
I can see standardization happening in early markets (where the vendors don’t have market power), or in areas where there is a critical open source software project that has the power to drive standards.
You also forget to mention that if Flikr is your photo album, they have all your data but are undoubtedly not liable for the loss of said data. Same goes for Gmail, Yahoo Mail, delicious, and other such web services. Thoughts like the above are why I run my own email/web/Jabber/etc servers on my own machine colocated with an ISP. Sure, I sometimes lack the latest and greatest features… but I’m sure that neither my data nor the services I host can go away at someone else’s whim.