There’s been some discussion about the EOLing of the Google SOAP APIs. I worked as the SOAP APIs product manager in 2002, with Nelson Minar and another Google engineer (both of whom did a great job on many Google projects while I was there).
Some blogs have criticized the move, but I wasn’t surprised. The SOAP APIs were very experimental when launched. They were never “finished”, but they served as a useful prototype for other APIs Google has created. If I were still at Google recommending products whose time had come to be EOL’d, the SOAP APIs would be on the list.
While the SOAP user base was steady, the product never exploded with demand. It’s somehow too complicated for the “long tail” of web designers (which the AJAX Search API seems to serve better), yet too limited for serious development.
Meanwhile, Amazon’s S3 seems to be taking the next step into a real business model for Web services. Dave Winer is right that there are opportunities here for other companies like Amazon. And entrepreneurs, too.
On the other hand, turning off new keys probably needed to happen without pre-announcement, lest a blogswarm lead to clever hackers registering 1 million API keys and selling them on eBay (or something similarly random).
However there’s no good way to EOL a product people are using – criticism was inevitable. Hopefully newer and better things will come.
What I’ll miss most? The cute little graphic of a robot searching with Google that was made for the launch.