Endedness

[rivest@theory.lcs.mit.edu (Ron Rivest)]

There are two issues here which we must not confuse.  One is the
mapping from the external representation to the interal
representation.  The other is the interpretation of the internal
representation as an integer.

SPKI/SDSI does not have "typed" data; it just has byte-strings.  

The mapping from external representation to internal (memory)
representation will not have access to any information about the kind
of data being represented by the byte string; it could be an integer,
a character string, or a computer program.  

My working assumptions are that memory is a byte array, and that the
first byte read from the external representation will go into the
lowest-indexed memory byte, and following bytes of the external
representation will go into successively higher-indexed memory bytes.

I think this is the only sensible way to do this, if you don't know the
length ahead of time.

The other issue is how to interpret the byte array as a bignum.  I suggested
that the low-indexed (first) byte be treated as the least-significant
byte, but Perry suggests that it be treated as the most-significant byte.

I am willing to accept the chair's suggestion on treating the
low-indexed byte as the most-significant byte of a bignum, but feel
strongly about the correspondence between external representations and
internal representations.  Phill Hallam-Baker's arguments make sense
to me there.

Is the standard "endians" argument about problem (1) or problem (2) above?

Ron Rivest

---

• Prev by Date: endian wars
• Next by Date: Re: Bignums
• Prev by thread: Re: endian wars
• Next by thread: Endian orders
• Index(es):
  • Main
  • Thread