GMail already
is a platform; look at Mailbox.app -- it's a new UI for the GMail "platform". Building on top of the GMail platform obviates the need to actually do any of the grungy mail stuff. But as one commenter noted elsewhere in this thread, the lock-in implications are pretty sobering.
We made Inky from scratch -- without building on top of GMail, Outlook, etc. -- partly to avoid the lock-in, but primarily because it's hard to imagine fundamentally changing the email user experience without touching any of the code for the mail platform itself. The downside, of course, is the usual trade-off: by not layering on top of somebody else's already-debugged stack, we have to make our own huge raft of code work right.
The analogy to Lotus Notes should be a wake-up call to CIOs everywhere; our analysis of the enterprise email/messaging space makes it clear that the big companies still using Lotus Notes are doing so primarily because they have too much investment in the Notes ecosystem ("intranet web sites" built on top of Notes, etc.) to switch. They come up with an internal estimate of the cost to convert all the stuff they've built on the Lotus platform over to something else and realize that figure exceeds the likely N-year savings from switching. So, conversion project not approved by CFO. Next!