Well, but what about after R1? I know it's too early to say, but I think this is a matter that we had better talk about from now. My main concern is that if BFS gets replaced by something else (I hope not), the data stored with it might be at risk. Even with the proper tools for data transfer, it could happen that not all the data will be safely transferred. As I pointed out in a previous message, my worry are the extended attributes, which in BFS are unlimited, while in other filesystems, as far as I know, are subject to limitations (in number and size). What would happen, for example, if I have a lot of extended attributes for a single file, and the target filesystem to which that file is being transferred can only support a limited number of attributes? Part of the stored metadata would then be lost. This is my worry. That's why I would be happier to stay with BFS, and see it being further developed, rather than moving to ZFS or to some other filesystem.
All the reasons you mentioned are reasons why BFS will likely remain central to Haiku (in revised form, most likely) well into the future. In the worst case, however, you can be assured that BFS support shall remain in the future for a considerable duration regardless of any default file system change, so your data will be accessible and usable perpetually.
It would also be exceedingly unlikely that Haiku would ever desire to move to a less robust file system considering the huge advantages engendered by the used of BFS. I find it more likely that a modified version of ZFS would be used and the migration of data would be seamless - as a worse case scenario...
--The loon