From the outside looking in, i can tell you that the adoption and willingness to upgrade to Exchage 2010 has been surprisingly rapid for us as well, typically customers won't roll out huge changes to their environments till an SP1 level has come out, many organizations still even run Exchange 2000 and Windows 2000 servers with only the lack of support being what forces them to move.
Historically if you look at the willingness to upgrade from 5.5 to 2000, or 2000 to 2003 etc the willingness to upgrade to 2010 is significantly higher than previous versions, what i will disagree with as well, is that Exchange 2010 whilst a big improvement is not that far above and beyond Exchange 2007 as that was the biiggest architectural changes (CAS/HUB/Transport etc all being introduced)
As for running EV2007 client (not EV7) with Outlook 2010, I could be wrong, but partial support is not a solution and for many customers won't be entertained, after all many companies such as the one i work for have to distribute these clients to tens of thousands of users which have to be tested, packaged, deployed etc which also takes considerable amount of time, we wouldnt want to have to distribute yet another client to all of our users.
But really it comes down to the way that Outlook 2010 has changed how the Outlook Add-Ins have to interface with Outlook 2010 by changing this to OOM.
Now the problem with this is, between Outlook 2007 and Outlook 2010, Microsoft have had three years to work on and develop those API's and hooks, however for developers, they have had limited time and limited documentation (which Microsoft is unfortunately infamous for) in which to be able to develop their outlook add-ins.
And again as mentioned when you consider the size of the client now, with Virtual Vault, Vault Cache etc, its not as simple as just changing a header and expecting everything to work, majorly significant development and even more QA has to be passed along to that, and each issue that is bought up by QA has to go back to Dev for a fix, and then more QA and regression testing be performed.
And also remember that EV has to be able to perform well with other new features added to outlook not seen in 2007 and 2003, Microsoft have the luxury of being able to change features, add features, change interfaces etc and take their time needed to perfect it, I think same courtesy needs to be extended towards developers who have to react and perform significant re-working of existing code.
As for the planning of the client, if you read the statement from Alex Brown it will be an EV9 SP1 client that introduces Outlook 2010 functionality, as for providing support for EV8, theres really no point considering EV9 clients with EV8 servers will most likely be compatible, so just roll out the EV9 client and that way if/when you upgrade to EV9 on the servers, you'll already have the functionality there