knightPhlight <ntech(_at_)ctive(_dot_)com> writes:
"640k should be more than enough memory for anybody." - Bill Gates
Cees de Groot <cg(_at_)pobox(_dot_)com> writes:
Some people didn't see a need for more than 640k, either...
Uncanny, same message simultaneously.
Anyway I side with knightPhlight -- 256 isn't enough, and 64k isn't
enough either. Who knows what you'll need. Make it a string. It'll
only cost you one byte (or nothing compared to Cees's call for 2
bytes).
As another thing what happened to the unix philosphy of human readable
file formats? Or at least file formats that you can reasonably pick
apart with a hex editor?
Why are PGP so keen to hand huffman encode everything?
(single bit fields, 1 byte identifier, variable length of length
fields etc.)
PRZ wrote PGP 1.0 for DOS first, perhaps this is why ... Personally I
would start over and use some thing that would cause pgp2.x to give up
gracefully as the first byte.
I can grant something for cleartext signatures perhaps, but other than
that it's hardly an issue.
If pgp are concerned about size at this nit-picking hand huffman
encoding level: they made the wrong choice with El Gamal, they
would've saved space with straight Diffie-Hellman, and not had to
generate so many random numbers as a bonus. (A pgp employee who I
won't name admitted in the past when this issue was raised, oh well,
it's only a few bytes).
Another thing while I'm at it:
The constant use of Diffie-Hellman to refer to El-Gamal is misleading.
Is seems to get referred to as "Diffie-Hellman" with occasional note
that actually they mean El-Gamal, "a variant of Diffie-Hellman."
Is this mis-nomenclature motivated by a belief that more people will
have heard of Diffie-Hellman and hence get a warm fuzzy feeling or
something?
It's El-Gamal, damn it!
(re hand huffman encoding fun -- it's something I have practice at, my
sig is an implementation of RSA in perl and dc, which has been hand
compressed in such a way that it still functions -- even gzip can't
wring much out of it now -- actualy it gets bigger if you consider
that it is already seven bit clean).
Adam
--
Have *you* exported RSA today? --> http://www.dcs.ex.ac.uk/~aba/rsa/
print pack"C*",split/\D+/,`echo "16iII*o\U(_at_){$/=$z;[(pop,pop,unpack"H*",<>
)]}\EsMsKsN0[lN*1lK[d2%Sa2/d0<X+d*lMLa^*lN%0]dsXx++lMlN/dsM0<J]dsJxp"|dc`