the debate is academic anyway. I highly doubt Google enjoys repeating failure, but that is what they might be walking into.
The whole idea here should be to solve a problem; this web-based OS architecture is a solution looking for a non-trivial problem. With the VAX-VMS for example, and other time-shared computer architectures (not quite what google is going for, but the concepts are pretty similar in many respects) the reason they did it, was simply because the computer itself was so large and expensive; they couldn't hope to have more then one, and certainly not an entire laboratory.
The very idea of a "personal" computer should be to emphasize the "personal" idea; in that respect the google OS will not be a "personal" computer operating system, since data is not stored on a personally owned hard drive, but rather on a hard drive on some server somewhere. The question here is What exactly does this achieve? Now all disk access to that data is slower. The main thing they vouch is that it will be accessible "anywhere"; but I think they need to re-examine just how trivial it is to copy files that are necessary onto a flash drive, or the large number of ways that this can be done within other operating systems; synchronization between a local and a ftp folder, for example.
did I mention that people "personal" data will be stored on hard drives owned by google? because personally I find that a bit disconcerting.
Another important thing is wether even google could create the appropriate architecture, and how long will they? People say that MS is controlling the hardware market; but really unless you want the latest Windows OS you have no need for updated hardware; old computers still run windows 98 fine, for example.
But if google succeeds in launching this platform and it becomes popular, google can make partnerships with hardware vendors and even force users to have certain hardware in order to run, or use certain features. Google already provides advertised links that may or may not be either relevant or safe, so it's no big revelation to realize that they are in fact a company and the bottom-line is really the very reason they do anything. But what exactly does releasing this OS do for their bottom line? They will need to purchase servers and a very powerful back-end for all this data that people store; I would hope they would have backups of that as well in case of failures, not to mention 24/7 monitoring staff. This is expensive. They wouldn't be doing this if they didn't think they could profit in the long run. And sure, it sounds kind of like a conspiracy theory; but I think what Google is really after is the data itself; the ability to mine personal data and provide the very same "google experience", like "personalized" advertisements; and even the automatic subscription to services, masked under the thin veil of some sort of silly google trademark. They already mine data from peoples browsing habits via Google AdSense; by strengthening both the scope of the display of advertisements as well as the amount of data to mine (that is, browser that may be even set up to block google adsense to a entire OS developed solely for the purpose of said data mining) it would vastly improve their abilities to mass-market other companies products to the proper demographics and to people interesting in such products. This means they could boost the pricing of said contracts.
Is this what they are doing? I have no idea; it's just a theory. But given what I've seen of Google, especially with regards to Chrome's early EULA's, it would seem that this has been their goal for some time.