Re: Wrong Endian platform suggestions

[Robert Elz <kre%munnari.OZ.AU@localhost>]

    Date:        Wed, 10 Aug 2016 09:27:51 +0200
    From:        Andreas Krey <a.krey%gmx.de@localhost>
    Message-ID:  <20160810072751.GD18740%inner.h.apk.li@localhost>

  | BigEndian looks best when used in protocols, yet there is
  | the argument that it is inconsistent in bit and byte addressing.

"bit addressing" is mostly a software convention - when I started out
with computing (on big endian systems) the convention was that the bits
were numbered left to right, just the same as everything else.  That is,
bit 0 was the most significant (or sign) bit.

With that convention, big endian is just as as consistent as little endian
is when bits are numbered so bit 0 is the least significant.

However, as bit numbering is (mostly) purely software we don't need to
implement both versions, and as almost all of the software we use was
developed on little endian systems originally (ie: vax), we conventionally
number all bits in little endian formats, even on big endian systems (when
I first encountered that it bothered me ... I really wanted to number bits
the other way!)

Endianness really is an argument (or discussion) that isn't worth
having - neither is "right", and there are advantages and
disadvantages to each (if that wasn't the case, one would have vanished
by now.)

kre