As Allan Sulkin regularly used to remind me, no one user uses all those features, but they're in the product because somebody needed each one. This uncomfortable reality tends to get brought to the attention, in a convincing way, of those newcomers who'd scoffed at the Big Feature List.
Now we move onto Unified Communications, which genuinely could render some of those PBX features obsolete as people find new ways and multiple channels for carrying on the communications that used to be restricted to the single voice channel in the past. But there's another thing about UC.
An emerging line of thought about UC deployment is that you don't need to--in fact, shouldn't--deploy every major UC application/media ubiquitously across all your end users. Not all of your workers need mobility, or conferencing, or whatever the case may be. So it's important to match the functionality with the user.
In this sense, then, isn't UC just that old PBX feature set writ large? The PBX feature set built up over time, as demanded features were added. The difference with UC is that, instead of waiting for the next software generic to be released with the new features, you add the new UC feature as a new application, most likely running on a new server and deployed to a discrete subset of users based on software licenses--rather than having the whole mass of features sitting on one big platform, theoretically available to everyone but, in practice, segmented by the individual user's communications needs.
UC ought to represent a more efficient model: Users don't have access to communications functions they don't need. On the other hand, that's already the case now. I don't have access to the features that I don't need on our office's key/hybrid system. My lack of access is enforced not by corporate policy or the lack of a software license, but by my own ignorance: Any feature I don't need, I never bothered to learn how to use.
The corollary ought to be that, under the new model, the enterprise saves money because it's only buying the aggregate amount of feature/functionality that its users will actually use. In practice, do you think that's how it's going to work?
Microsoft has made a big deal about the cost of voice software falling drastically over the coming years. But we can see in the retort by Siemens and others why my previous paragraph seems unlikely: Mark Straton and others at Siemens have pointed out that, even as software prices fall, the costs for integration and management are likely to go up. Ask yourself: Which is easier to manage: 450 features on a single platform, or dozens (or hundreds) of servers running a wide variety of applications?
I'm not knocking the new vision; it's a lot more functional, it incorporates a lot richer range of communications, and it could offer a lot of business benefits. But let's not kid ourselves that chucking the PBX and its 450 features will make things any simpler or cheaper.