Is Google's Proprietary Tech Stack Destroying Its Acquisitions?

While Google has bought plenty of small startups, almost none of those deals have amounted to very much. It almost seems like most of the startups disappear into Google forever. There are a few exceptions such as YouTube and (maybe) Writely. But the list of startups that have simply languished or died is much longer. TechCrunchIT is running an interesting post that suggests one of the key reasons: Google's proprietary tech stack. While Google is a big open source supporter for lower level infrastructure, once you get above that -- it's very much a strong believer in doing everything its own way. I've heard from friends at Google about the difficulty they've had learning to deal with Google's tech stack -- and certainly have heard how it's slowed down the progress of some Google acquisitions while they learn how to "transition."
In fact, some have pointed out that this is one of the side benefits to Google's AppEngine offering. Since it exposes some of Google's tech stack to folks for them to develop and run their applications, it will make it much easier to integrate them into Google at a later date. So, for startups whose strategy is to get acquired by Google (and, I should note, if you start with that strategy, you're probably going to fail), it may make sense to develop on AppEngine just because you're already signaling to Google that the integration costs are significantly lower.
Still, this highlights one of the major downsides to Google's belief that it can do everything much better than everyone else by starting from scratch: in doing so, it actually makes it much harder to capitalize on synergies from many acquisition targets. Yes, there are reasons to go against the "standard" way of doing things, but there are significant costs as well.