The self-fulfilling prophecy[1] is a powerful phenomenon. If Google pledges 3 years support, then I will assume it's going to be killed in 3 years and not even bother trying to build a business on it, even though Google only means that as a minimum. If everybody assumes the same, then the API probably will end up dead in 3 years, because nobody is using it. On the other hand, since Google makes no such pledge, I don't feel like this API has an expiration date. Instead, I look at this as Google's latest product, and I might want to be the first to integrate it into my service, because it could be the next big thing. If enough people think that way, then the API will thrive and stick around. Sure, the API could still disappear, but any startup built on it has a greater chance of dying before the API does.
Google's problem right now is that it has pissed off a vocal minority with the Google Reader shutdown, so any announcement is being met with pessimism and mistrust.
No expiration date is just an evil omission. The risk is of obsolescence is there, whether they state the date or not. Making the risk visible to everyone makes the situation clearer for everyone. All that is needed is a table like http://support.microsoft.com/lifecycle/?ln=en-gb&c2=1173
Google's problem right now is that it has pissed off a vocal minority with the Google Reader shutdown, so any announcement is being met with pessimism and mistrust.
1: http://en.wikipedia.org/wiki/Self-fulfilling_prophecy