Cryptography-Digest Digest #509, Volume #14 Sun, 3 Jun 01 21:13:00 EDT
Contents:
Re: HELP WITH RSA ENCRYPTION/DECRYPTION INCLUDING GARNER CRT ALGORITHM ("Jeffrey
Walton")
Re: BBS implementation ("Tom St Denis")
Re: PRP vs PRF (was Luby-Rackoff Theorems) ("Scott Fluhrer")
Re: BBS implementation ("Tom St Denis")
Re: Help with Comparison Of Complexity of Discrete Logs, Knapsack, and Large Primes
(sisi jojo)
Re: PRP vs PRF (was Luby-Rackoff Theorems) ("Tom St Denis")
Re: BigNum Question ("Jeffrey Walton")
Re: BigNum Question ("Tom St Denis")
Re: PRP vs PRF (was Luby-Rackoff Theorems) ("Scott Fluhrer")
Re: Help with Comparison Of Complexity of Discrete Logs, Knapsack, and Large Primes
("Tom St Denis")
Re: BigNum Question ("Jeffrey Walton")
Re: BigNum Question ("Tom St Denis")
Re: PRP vs PRF (was Luby-Rackoff Theorems) ("Tom St Denis")
Re: BigNum Question ("Jeffrey Walton")
Re: BigNum Question ("Tom St Denis")
----------------------------------------------------------------------------
Reply-To: "Jeffrey Walton" <[EMAIL PROTECTED]>
From: "Jeffrey Walton" <[EMAIL PROTECTED]>
Subject: Re: HELP WITH RSA ENCRYPTION/DECRYPTION INCLUDING GARNER CRT ALGORITHM
Date: Sun, 3 Jun 2001 20:14:25 -0400
I agree basically with Mr. Ashwood and Mr. Wooding. One thing I would
probably say is to use small, easy numbers to work out the program
first. This way, results can be checked with a hand calculator.
For example, 065^17 = 6599743590836592050933837890625. This is too big
a number to use to debug your implementation.
Can you choose smaller N, e, and d so that these can be worked out by
hand (and don't worry about strong primes at this point)? Also, rather
than encode a string (ABCD), start with 1 character at a time (A).
Jeff
"Sam" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
May be I am missing something or taking a wrong approach.
Plaintext 'm' to be encrypted : m= ABCD12345
* convert chars in 'm' to a byte array with ascii values of the chars in
the plaintext :
inputStr = [065 066 067 068 049 050 051 052 053]
encryption ciphertext c = m^e mod n
where e =17, and
n (size 128 bits) =
604617013676588903089819459843928561595967634140770680278581402097188044
430156586545307409371288684871273873727250749538931794090167705319066304
181984583
so c will be a string (065^e mod n 066^e mod n 067^e mod 068^e modn
049^e modn 050^e mod n 051^e mod n 052^e mod n 053^e mod n), whic
may come out some thing like this
CipherText:
!,0129140163,16926659444736,762939453125,0129140163,16926659444736,16926
659444736,0232630513987207,16926659444736
,232630513987207,0232630513987207,16926659444736,2251799813685248,012914
0163,17179869184,16677181699666569,0129140163,7629394
53125,0129140163,0129140163,762939453125,1,0129140163,762939453125,13107
2,02251799813685248,762939453125,129140163
when this ciphertext is passed to decrypt function to get original
messege m
m = c^d mod n
where d =
exp
106697120060574512309968139972457981458111935436606590637396718017150831
370018871731047547545813561329601322725228331964181118052002625130275185
842812553
m =(0129140163^d mod n, 16926659444736^d mod n, ............
................... ....................... , 762939453125^d mod
n,129140163^d mod n)
= 065 066 067 068 049 050 051 052 053 which when converted to char gives
back m = ABCD12345.
IS THIS THE CORRECT WAY BUT WHEN USING GARNER ALGORITHM TO DECRYPT
I HAVE PROBLEMS
BigInteger temp = p.subtract(CONST_ONE);
dP = e.modInverse(temp);
System.out.println("dP: " + dP);
dQ = e.modInverse(q.subtract(CONST_ONE));
System.out.println("dQ: " + dQ);
m1 = c.modPow(dP,p);
System.out.println("m1: " + m1);
m2 = c.modPow(dQ,q);
System.out.println("m2: " + m2);
BigInteger qInv = q.modInverse(p);
System.out.println("qInv: " + qInv);
h = m1.subtract(m2);
h = h.multiply(qInv);
h = h.mod(p);
System.out.println("h: " + h);
m = m2.add(q.multiply(h));
System.out.println("M: " + m.toString());
return m;
* Do I pass ciphertext string
CipherText:
!,0129140163,16926659444736,762939453125,0129140163,16926659444736,16926
659444736,0232630513987207,16926659444736
,232630513987207,0232630513987207,16926659444736,2251799813685248,012914
0163,17179869184,16677181699666569,0129140163,7629394
53125,0129140163,0129140163,762939453125,1,0129140163,762939453125,13107
2,02251799813685248,762939453125,129140163
to both
m1 = c.modPow(dP,p);
m2 = c.modPow(dQ,q);
Then I will get two different byte arrays
then how do I proceed to get m
BigInteger qInv = q.modInverse(p);
h = m1.subtract(m2);
h = h.multiply(qInv);
h = h.mod(p);
m = m2.add(q.multiply(h));
Will I get to right answer, will I get real impimprovement in efficiency
using Garner algorithm. does somebody have some comparision results.
messege may be much bigger than ABCD12345
can somebody give me some example
Thanks a lot in advance
Sam
------------------------------
From: "Tom St Denis" <[EMAIL PROTECTED]>
Subject: Re: BBS implementation
Date: Mon, 04 Jun 2001 00:13:11 GMT
"lcs Mixmaster Remailer" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> [EMAIL PROTECTED] writes, regarding using a random
> Blum-Blum-Shub starting point:
> > How do I know it is not on a short or degenerate cycle?
>
> If it is, then your number can be easily factored. Since you are probably
> comfortable assuming that random RSA moduli can't be easily factored,
> you must be equally comfortable assuming that a random BBS starting
> point is not on a short cycle.
I know I should know this but how doest the a short cycle help you factor?
i.e let's say with F : x => x^2 mod N, you get A,B,C,A as the outputs (i.e
cycle of period 3). (i.e B = A^2, C = A^4)
Oh is that it? If you know the period you know the power? i.e period k
gives you A^(2(k-1)) mod N = A, where 2k - 2 is the order of the
multiplicative group?
Tom
------------------------------
From: "Scott Fluhrer" <[EMAIL PROTECTED]>
Subject: Re: PRP vs PRF (was Luby-Rackoff Theorems)
Date: Sun, 3 Jun 2001 17:00:29 -0700
Tom St Denis <[EMAIL PROTECTED]> wrote in message
news:3_yS6.20233$[EMAIL PROTECTED]...
>
> "Scott Fluhrer" <[EMAIL PROTECTED]> wrote in message
> news:9fedm7$ke9$[EMAIL PROTECTED]...
> >
> > Tom St Denis <[EMAIL PROTECTED]> wrote in message
> > news:SKxS6.19715$[EMAIL PROTECTED]...
> > > A PRF is a pseudo-random function and a PRP is a pseudo-random
> > permutation?
> > > (I'm reading the paper Wagner posted today).
> > >
> > > So basically a PRF is any type of mapping and a PRP is any injection?
> > I think you have the right idea (you mean bijection, not injection, and
> > domain and range are the same), but lets get it pedantically accurate:
> >
> > A PRF is a distribution of functions generated from a particular source,
> > which has the following property: given a function F, which is either
> chosen
> > according to the distribution of the PRF, or a function chosen uniformly
> > randomly from the set of all functions with the same domain/range, it is
> > impossible (with some computational and query ceiling) to distinguish
> which
> > that particular F came from with probability 0.5 + \epsilon.
>
> So we say {x, y, ..., z} is a PRF if you can't tell how it was made from
F?
>
> > A PRP is exactly the same, except that we're dealing with permutations
and
> > not functions.
>
> I don't get this. Isn't a PRP a function?
It is, but it has considerably stricter requirements: the domain and the
range are the same, and (more importantly) that it be a bijection.
>
> > Quite often, we talk about a single function being a PRF, which is
> > convienent, and has an intuitive meaning (the function "acts randomly"
on
> > any input that hasn't been queried), but really doesn't have any precise
> > mathematical meaning -- it is analogous to the question of whether 42 is
a
> > random number.
>
> I will have to re-read this... thanks for the reply.
>
> Tom
>
>
------------------------------
From: "Tom St Denis" <[EMAIL PROTECTED]>
Subject: Re: BBS implementation
Date: Mon, 04 Jun 2001 00:15:45 GMT
"Tom St Denis" <[EMAIL PROTECTED]> wrote in message
news:roAS6.20968$[EMAIL PROTECTED]...
>
> "lcs Mixmaster Remailer" <[EMAIL PROTECTED]> wrote in message
> news:[EMAIL PROTECTED]...
> > [EMAIL PROTECTED] writes, regarding using a random
> > Blum-Blum-Shub starting point:
> > > How do I know it is not on a short or degenerate cycle?
> >
> > If it is, then your number can be easily factored. Since you are
probably
> > comfortable assuming that random RSA moduli can't be easily factored,
> > you must be equally comfortable assuming that a random BBS starting
> > point is not on a short cycle.
>
> I know I should know this but how doest the a short cycle help you factor?
>
> i.e let's say with F : x => x^2 mod N, you get A,B,C,A as the outputs
(i.e
> cycle of period 3). (i.e B = A^2, C = A^4)
>
> Oh is that it? If you know the period you know the power? i.e period k
> gives you A^(2(k-1)) mod N = A, where 2k - 2 is the order of the
> multiplicative group?
Correction 2k-2 would be a factor of the order, but not nessesarily the
order it self...
i.e A=1, B=1^2, C=1^4, etc... the period is 1 in this case...
Tom
------------------------------
From: [EMAIL PROTECTED] (sisi jojo)
Subject: Re: Help with Comparison Of Complexity of Discrete Logs, Knapsack, and Large
Primes
Date: 3 Jun 2001 17:30:08 -0700
"Tom St Denis" <[EMAIL PROTECTED]> wrote in message
news:<TNcS6.11869$[EMAIL PROTECTED]>...
> Um as far as I know if you learn the NFS and some basic number theory you
> should be able to factor quite easily. I know this is the case for QS (ok
> let's assume the size of the composite is trivial_
>
> Tom
You can't win $200,000 with QS and NFS, can you? :)
Actually the math in cryptography is not that hard, but the mathematical
language is very cryptic. If someone just re-writes all the papers in human
language, even teenagers can become good cryptographers. Otherwise, you need
years of schooling just to understand the math notations.
--Sisi
------------------------------
From: "Tom St Denis" <[EMAIL PROTECTED]>
Subject: Re: PRP vs PRF (was Luby-Rackoff Theorems)
Date: Mon, 04 Jun 2001 00:30:26 GMT
"Scott Fluhrer" <[EMAIL PROTECTED]> wrote in message
news:9fejsp$r2j$[EMAIL PROTECTED]...
>
> Tom St Denis <[EMAIL PROTECTED]> wrote in message
> news:3_yS6.20233$[EMAIL PROTECTED]...
> >
> > "Scott Fluhrer" <[EMAIL PROTECTED]> wrote in message
> > news:9fedm7$ke9$[EMAIL PROTECTED]...
> > >
> > > Tom St Denis <[EMAIL PROTECTED]> wrote in message
> > > news:SKxS6.19715$[EMAIL PROTECTED]...
> > > > A PRF is a pseudo-random function and a PRP is a pseudo-random
> > > permutation?
> > > > (I'm reading the paper Wagner posted today).
> > > >
> > > > So basically a PRF is any type of mapping and a PRP is any
injection?
> > > I think you have the right idea (you mean bijection, not injection,
and
> > > domain and range are the same), but lets get it pedantically accurate:
> > >
> > > A PRF is a distribution of functions generated from a particular
source,
> > > which has the following property: given a function F, which is either
> > chosen
> > > according to the distribution of the PRF, or a function chosen
uniformly
> > > randomly from the set of all functions with the same domain/range, it
is
> > > impossible (with some computational and query ceiling) to distinguish
> > which
> > > that particular F came from with probability 0.5 + \epsilon.
> >
> > So we say {x, y, ..., z} is a PRF if you can't tell how it was made from
> F?
> >
> > > A PRP is exactly the same, except that we're dealing with permutations
> and
> > > not functions.
> >
> > I don't get this. Isn't a PRP a function?
>
> It is, but it has considerably stricter requirements: the domain and the
> range are the same, and (more importantly) that it be a bijection.
Observation. Are PRFs a subset of PRPs?
Tom
------------------------------
Reply-To: "Jeffrey Walton" <[EMAIL PROTECTED]>
From: "Jeffrey Walton" <[EMAIL PROTECTED]>
Subject: Re: BigNum Question
Date: Sun, 3 Jun 2001 20:35:58 -0400
Try http://www.shoup.net/ntl/doc/tour-win.html
The author states you should be able to compile for a Mac. I D/L the
Windows version, but was never able to compile.
Its been my experience that if you are not developing on a *nix flavor,
you're basically out of luck for a big integer/big number package. What
a shame a basic package is not included with C++ (like the STL).
Jeff
"George" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
: I'm trying to develop a program for Macintosh and I need to operate on
: very large numbers. What is the best BigNum library for Macintosh
where
: the source code is also available?
:
: -George
: [EMAIL PROTECTED]
:
------------------------------
From: "Tom St Denis" <[EMAIL PROTECTED]>
Subject: Re: BigNum Question
Date: Mon, 04 Jun 2001 00:33:39 GMT
"Jeffrey Walton" <[EMAIL PROTECTED]> wrote in message
news:3b1ad716$0$[EMAIL PROTECTED]...
> Try http://www.shoup.net/ntl/doc/tour-win.html
>
> The author states you should be able to compile for a Mac. I D/L the
> Windows version, but was never able to compile.
>
> Its been my experience that if you are not developing on a *nix flavor,
> you're basically out of luck for a big integer/big number package. What
> a shame a basic package is not included with C++ (like the STL).
Not true. both GMP and MPI work out of the box (well MPI is easier) with
DJGPP. MPI also works with MS VC... MPI is designed to be very portable.
Tom
------------------------------
From: "Scott Fluhrer" <[EMAIL PROTECTED]>
Subject: Re: PRP vs PRF (was Luby-Rackoff Theorems)
Date: Sun, 3 Jun 2001 17:21:35 -0700
Tom St Denis <[EMAIL PROTECTED]> wrote in message
news:CEAS6.21014$[EMAIL PROTECTED]...
>
> "Scott Fluhrer" <[EMAIL PROTECTED]> wrote in message
> news:9fejsp$r2j$[EMAIL PROTECTED]...
> >
> > Tom St Denis <[EMAIL PROTECTED]> wrote in message
> > news:3_yS6.20233$[EMAIL PROTECTED]...
> > >
> > > "Scott Fluhrer" <[EMAIL PROTECTED]> wrote in message
> > > news:9fedm7$ke9$[EMAIL PROTECTED]...
> > > >
> > > > Tom St Denis <[EMAIL PROTECTED]> wrote in message
> > > > news:SKxS6.19715$[EMAIL PROTECTED]...
> > > > > A PRF is a pseudo-random function and a PRP is a pseudo-random
> > > > permutation?
> > > > > (I'm reading the paper Wagner posted today).
> > > > >
> > > > > So basically a PRF is any type of mapping and a PRP is any
> injection?
> > > > I think you have the right idea (you mean bijection, not injection,
> and
> > > > domain and range are the same), but lets get it pedantically
accurate:
> > > >
> > > > A PRF is a distribution of functions generated from a particular
> source,
> > > > which has the following property: given a function F, which is
either
> > > chosen
> > > > according to the distribution of the PRF, or a function chosen
> uniformly
> > > > randomly from the set of all functions with the same domain/range,
it
> is
> > > > impossible (with some computational and query ceiling) to
distinguish
> > > which
> > > > that particular F came from with probability 0.5 + \epsilon.
> > >
> > > So we say {x, y, ..., z} is a PRF if you can't tell how it was made
from
> > F?
> > >
> > > > A PRP is exactly the same, except that we're dealing with
permutations
> > and
> > > > not functions.
> > >
> > > I don't get this. Isn't a PRP a function?
> >
> > It is, but it has considerably stricter requirements: the domain and the
> > range are the same, and (more importantly) that it be a bijection.
>
> Observation. Are PRFs a subset of PRPs?
Other way around. All permutations are functions, but not all functions are
permutations. In addition, unless the query ceiling is considerable less
than sqrt(|range|), or if |range| is quite small, a PRF has an extremely
small (but nonzero) probability of being a permutation.
--
poncho
------------------------------
From: "Tom St Denis" <[EMAIL PROTECTED]>
Subject: Re: Help with Comparison Of Complexity of Discrete Logs, Knapsack, and Large
Primes
Date: Mon, 04 Jun 2001 00:35:10 GMT
"sisi jojo" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> "Tom St Denis" <[EMAIL PROTECTED]> wrote in message
news:<TNcS6.11869$[EMAIL PROTECTED]>...
> > Um as far as I know if you learn the NFS and some basic number theory
you
> > should be able to factor quite easily. I know this is the case for QS
(ok
> > let's assume the size of the composite is trivial_
> >
> > Tom
>
> You can't win $200,000 with QS and NFS, can you? :)
>
> Actually the math in cryptography is not that hard, but the mathematical
> language is very cryptic. If someone just re-writes all the papers in
human
> language, even teenagers can become good cryptographers. Otherwise, you
need
> years of schooling just to understand the math notations.
I agree,but sometimes the notation helps the discussion go quicker.
My peeve is with group notation. Too many slashes and such... arrg!
And no, the avg teen would not be a good cryptographer. I would say the avg
high school grad *could* be a good cryptographer. But a 15 yr old typically
doesn't have the requisit math knowledge to design/analyze anything.
Tom
------------------------------
Reply-To: "Jeffrey Walton" <[EMAIL PROTECTED]>
From: "Jeffrey Walton" <[EMAIL PROTECTED]>
Subject: Re: BigNum Question
Date: Sun, 3 Jun 2001 20:46:59 -0400
Thanks Tom (I know you coached me on this previously).
I'd prefer not to run DJGPP. I will look into MPI (searching for it as
we write).
Jeff
"Tom St Denis" <[EMAIL PROTECTED]> wrote in message
news:DHAS6.21026$[EMAIL PROTECTED]...
:
: "Jeffrey Walton" <[EMAIL PROTECTED]> wrote in message
: news:3b1ad716$0$[EMAIL PROTECTED]...
: > Try http://www.shoup.net/ntl/doc/tour-win.html
: >
: > The author states you should be able to compile for a Mac. I D/L
the
: > Windows version, but was never able to compile.
: >
: > Its been my experience that if you are not developing on a *nix
flavor,
: > you're basically out of luck for a big integer/big number package.
What
: > a shame a basic package is not included with C++ (like the STL).
:
: Not true. both GMP and MPI work out of the box (well MPI is easier)
with
: DJGPP. MPI also works with MS VC... MPI is designed to be very
portable.
:
: Tom
:
:
------------------------------
From: "Tom St Denis" <[EMAIL PROTECTED]>
Subject: Re: BigNum Question
Date: Mon, 04 Jun 2001 00:48:51 GMT
"Jeffrey Walton" <[EMAIL PROTECTED]> wrote in message
news:3b1ad9aa$0$[EMAIL PROTECTED]...
> Thanks Tom (I know you coached me on this previously).
>
> I'd prefer not to run DJGPP. I will look into MPI (searching for it as
> we write).
Yup. MPI is keen. You only need the three files mpi.c, mpi.h and
mpi-config.h from it. No complicated makefiles or autoconf scripts. I've
imported it into MSVC in like 2 seconds...
I've used it in Linux and DOS with the same code ... etc..
it's reasonably fast too. Not blazing fast but good enough for most tasks
Tom
------------------------------
From: "Tom St Denis" <[EMAIL PROTECTED]>
Subject: Re: PRP vs PRF (was Luby-Rackoff Theorems)
Date: Mon, 04 Jun 2001 00:50:02 GMT
"Scott Fluhrer" <[EMAIL PROTECTED]> wrote in message
news:9fel2u$gt2$[EMAIL PROTECTED]...
>
> Tom St Denis <[EMAIL PROTECTED]> wrote in message
> news:CEAS6.21014$[EMAIL PROTECTED]...
> >
> > "Scott Fluhrer" <[EMAIL PROTECTED]> wrote in message
> > news:9fejsp$r2j$[EMAIL PROTECTED]...
> > >
> > > Tom St Denis <[EMAIL PROTECTED]> wrote in message
> > > news:3_yS6.20233$[EMAIL PROTECTED]...
> > > >
> > > > "Scott Fluhrer" <[EMAIL PROTECTED]> wrote in message
> > > > news:9fedm7$ke9$[EMAIL PROTECTED]...
> > > > >
> > > > > Tom St Denis <[EMAIL PROTECTED]> wrote in message
> > > > > news:SKxS6.19715$[EMAIL PROTECTED]...
> > > > > > A PRF is a pseudo-random function and a PRP is a pseudo-random
> > > > > permutation?
> > > > > > (I'm reading the paper Wagner posted today).
> > > > > >
> > > > > > So basically a PRF is any type of mapping and a PRP is any
> > injection?
> > > > > I think you have the right idea (you mean bijection, not
injection,
> > and
> > > > > domain and range are the same), but lets get it pedantically
> accurate:
> > > > >
> > > > > A PRF is a distribution of functions generated from a particular
> > source,
> > > > > which has the following property: given a function F, which is
> either
> > > > chosen
> > > > > according to the distribution of the PRF, or a function chosen
> > uniformly
> > > > > randomly from the set of all functions with the same domain/range,
> it
> > is
> > > > > impossible (with some computational and query ceiling) to
> distinguish
> > > > which
> > > > > that particular F came from with probability 0.5 + \epsilon.
> > > >
> > > > So we say {x, y, ..., z} is a PRF if you can't tell how it was made
> from
> > > F?
> > > >
> > > > > A PRP is exactly the same, except that we're dealing with
> permutations
> > > and
> > > > > not functions.
> > > >
> > > > I don't get this. Isn't a PRP a function?
> > >
> > > It is, but it has considerably stricter requirements: the domain and
the
> > > range are the same, and (more importantly) that it be a bijection.
> >
> > Observation. Are PRFs a subset of PRPs?
>
> Other way around. All permutations are functions, but not all functions
are
> permutations. In addition, unless the query ceiling is considerable less
> than sqrt(|range|), or if |range| is quite small, a PRF has an extremely
> small (but nonzero) probability of being a permutation.
Oh yeah that makes sense. So not all PRFs are invertable (i.e one-to-one)
right?
Sorry if these questions seem lame. They don't teach this much group theory
(or whatever this is) in school...
Tom
------------------------------
Reply-To: "Jeffrey Walton" <[EMAIL PROTECTED]>
From: "Jeffrey Walton" <[EMAIL PROTECTED]>
Subject: Re: BigNum Question
Date: Sun, 3 Jun 2001 21:01:39 -0400
:
: it's reasonably fast too. Not blazing fast but good enough for most
tasks
:
At this point, I don't need speed - just a clean and easy to use
implementation.
I don't agree with Gossling (Emacs is an IDE). I prefer my beloved
Microsoft Visual C++ with intellisense, autocomplete, an no more
makefiles!
Jeff
"Tom St Denis" <[EMAIL PROTECTED]> wrote in message
news:TVAS6.21051$[EMAIL PROTECTED]...
:
: "Jeffrey Walton" <[EMAIL PROTECTED]> wrote in message
: news:3b1ad9aa$0$[EMAIL PROTECTED]...
: > Thanks Tom (I know you coached me on this previously).
: >
: > I'd prefer not to run DJGPP. I will look into MPI (searching for it
as
: > we write).
:
: Yup. MPI is keen. You only need the three files mpi.c, mpi.h and
: mpi-config.h from it. No complicated makefiles or autoconf scripts.
I've
: imported it into MSVC in like 2 seconds...
:
: I've used it in Linux and DOS with the same code ... etc..
:
: it's reasonably fast too. Not blazing fast but good enough for most
tasks
:
: Tom
:
:
------------------------------
From: "Tom St Denis" <[EMAIL PROTECTED]>
Subject: Re: BigNum Question
Date: Mon, 04 Jun 2001 01:01:22 GMT
"Jeffrey Walton" <[EMAIL PROTECTED]> wrote in message
news:3b1add1a$0$[EMAIL PROTECTED]...
> :
> : it's reasonably fast too. Not blazing fast but good enough for most
> tasks
> :
>
> At this point, I don't need speed - just a clean and easy to use
> implementation.
>
> I don't agree with Gossling (Emacs is an IDE). I prefer my beloved
> Microsoft Visual C++ with intellisense, autocomplete, an no more
> makefiles!
Well I like makefiles, but if you have ever tried to build a linux or unix
specific "portable" package with DJGPP you will encounter nightmares. They
use "autoconf" and other stupid scripts...
I like MPI because even if it did have a makefile it would be trivial and
should build out of the box for anything... no stupid config scripts etc..
(Note: I got this from trying to build aspell....arrg)
Tom
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list by posting to sci.crypt.
End of Cryptography-Digest Digest
******************************
