Thus spake Bill Studenmund ("BS> ") sometime Today...

BS> As for partitions, if we haven't done wedges, sd0a, sd0b, and so on would
BS> (I expect) show up as "scsibusX target Y lun Z partitionP" where P is the
BS> partition number in whatever the kernel reads as a disklabel. If we have
BS> wedges, only the raw disk would show up (sd0, sd1, etc.) and all the other
BS> partitions would be done by the wedge meister (whatever is doing wedges).
BS> The wedge meister would be doing the same thing, except it would be fining
BS> partitions with more descriptive locators, like "mbr1" for partition 1 in
BS> the main mbr, or "nbsd4" for partition 'e' in a NetBSD disklabel, or maybe
BS> "mbr2 nbsd4" for partition 'e' in a NetBSD partition in mbr slot 2.

Hm.  It's a matter of change, I'm sure, but I'm rather attached to the
"/dev/r?<type><ord><partition>" notation -- it's quick to type.  I don't
know what "wedges" would entail, but seeing as the tendency hovers between
having one huge partition and slicing up the disk like a large pizza,
I don't see a huge gain.  But then, I don't understand the concept of
"wedges" just yet, either.

Thanks for bearing with me.

				--*greywolf;
--
NetBSD: Someday, we won't burn your toast.