At 04:58 AM 7/21/00 -0800, Marco.Cimarosti@icl.com wrote:
>If UCS-2LE is a *standard* encoding (and it is in fact mentioned in UTR-17),
>how does VFAT directories qualify as a "higher level protocol"?
>My understanding of "higher level protocol" is that it is a *non* standard
>usage of some kind, allowed internally (or within a private group), that
>should not be transmitted to the world at large.
>Does this mean that MS's VFAT directories miss something (e.g. a BOM) to be
>a true UCS-2LE?
>Or are you simply meaning that the internals of an operating system are a
>"higher level protocol" by definition (even if they casually comply with
Since you have to abide by other specifications in order to even get to the
data field with the directory name, all of VFAT is a higher level protocol.
In that sense, any OS and its API are higher level protocols.
Now, HLP's are definitely free to override certain things (including byte
order), definitely not free to override others (no use of unassigned code
points) and there's (in my view) a bit of a gray area in between where our
understanding of where Unicode needs to allow flexibility for
implementations sake and where it must restrict implementations in order to
make meaningful guarantees about what constitutes Unicode is likely to
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:21:06 EDT