I think that in the next version of LR, it is extremely important to have a fully supported parametric 3rd party plugins possibility:
A better plugin API would make LR a more attractive platform for developers, but it doesn't seem to tie in with Adobe's strategy. They've been very aggressive in incorporating what were previously plug-ins in their programs. This includes HDR and stitching in PS, and noise reduction, sharpening and lens correction in LR. I think the main reason why Adobe is doing this is to add and update features to encourage customers to upgrade to the latest version of their
software to capture the revenue: if a customer buys or upgrades an HDR plugin the money goes to the manufacturer of the plugin, not Adobe.
Also, I wonder if the attach rate of commercial plugins is significant enough to warrant the expense in improving the API - ie. if the attach rate is 5%, an improved API might improve LR sales by only 5% x [multiplier for the impact of the improved API], which is likely much smaller than 5% unless it's a killer plugin which drive attach rates.
I don't know if the SDK and use of API costs anything in fees or royalties, so depending on that it will change the equation.
I'm sure NIK, OnOne, Picture Code, etc. are painfully aware of Adobe's strategy, and are wary of spending too much R&D on yet another Adobe program which will likely integrate their product in the future. Then again, many of those companies rely entirely on the existence of PS/LR, and have no choice other than reap the benefits of first-mover advantage.
In the end, quite a complicated question, and I'm ignoring the technical challenges (read: cost) in developing a more integrated API/SDK which I know nothing about. It is likely in the interest of Adobe to keep the SDK/API crippled to en extent to allow them to roll out their own fully integrated implementation of a feature to keep users upgrading to the newest version.