Locking in? What locking in? LR gives you the ability to easily export a raw file as a DNG to a new directory or you can directly access the original file from LR by hitting Command/Ctrl+R. If that's not good enough, then you can just browse to the directory containing the file just like you would have to do anyway with another converter.
That's clearly not good enough for me.
You have spent time inputing meta-data in LR to make it easier to query your RAW files, and now you'll have to browse directories or manually export a DNG each time you want to work with a different RAW converter? That's simply not manageable for those working with large amount of data with other converters. And I am not even discussing the problem that no-one supports DNG but Adobe.
All I am asking is the possibility to launch a different RAW converter from inside LR once it has located a file. It doesn't even have to be integrated in the LR interface.
How the links between the files are managed would have to be thought about, and Adobe alone would probably not be able to come up with something without cooperating with the other RAW converter software editors, which is why I suggested the creation of a plug-in mechanism. A plug-in mechanism is nothing but a simple set of standards that tell other software how they have to handle data so that the calling application can keep mastery of the workflow.
In the case of LR, all that would be needed is a simple way for LR to keep track of what files were generated by the external RAW converter based on the RAW file that was passed to it. The storage of the edits in the external raw converted would remain managed by the external converter (typically by storing a .xml file in the same directory as the RAW file).
For me, the ideal workflow would be a 4 + 2 steps mechanism inspired from RSP:
1. Browse images based on LR generated previews (new images after a shoot) or queries (older images you look for),
2. Apply simple meta data to the selected winners in LR (ranging from a simple Flag to more complex batch applied meta data),
3. Do a final conversion of the winning images in any RAW converter I like (DxO, Capture NX, Bibble, Silkpix,...),
4. Updating pointers in the LR database to the converted images. This would be done by LR based on the plug-in mechanism,
5. Open the converted image in PS or any other image editing software (noise reduction, sharpening,...),
6. Save back the layered image,
7. Use other LR functions like web galleries creations,...
I could even write specs for this. I'd have to be paid for doing that though.
Adobe is of course free to focus on an ACR centered workflow, but they will lose a unique chance to become universal and to occupy a truely central role in many photographers' workflow.
Anything can be done in software and Adobe has both the resources and skills to deal with the additional complexity. It can be V2, all that I need is a commited roadmap.