I think we're missing the point. What we _need_ is dictated by what we've
_got_. We have partitions. We don't have slices. So we use 'p'. If slices
get added into Linux one day then the device naming will have to be changed
anyway.
This is why I've pushed for using 'p' not 's'. It's more than aesthetics
or personal preference: 's' is WRONG for 'fdisk' generated disk layout.
You're forgetting that someone researched this and reported that "s" and "d"
originally stood for "device" and "subdevice", which is more general than
partitions or slices and not inherently wrong for fdisk. That's the definition
I adopted, not "slice". I still think that /dev/sd/c0b0t0d0 is a reasonable
choice for a whole disk and /dev/sd/c0b0t0d0s1 etc for PC style partitions. If
we need to incorporate further subdivisions for slices, we can easily extend
this to /dev/sd/c0b0t0d0s1.2 to refer to slice 2 within PC partition 1. I
think we should retain the convention of referring to logical partitions as
subdevices 5..N for compatibility with present Linux systems.
Leonard