From: ams@nospicedham.ludd.luth.se   
      
   Terje Mathisen  wrote:   
   > wolfgang kern wrote:   
   > > Thanks, I thought real needs caused this restrictions.   
   > > So my rules for such an Base64 decoder would allow ASCII 33..127 and   
   > > nothing less than 386 code.   
   > If you have 95 available characters you should get at least 6.5   
   > bits/char, i.e. with such a large alphabet you need to use something   
   > more efficient than Base64.   
      
   Actually, didn't this start out with any encoding allowed? Just that   
   it was some (subset of) ASCII. The actual encoding of the payload   
   could be anything.   
      
   I went to base64 because my first decoding of my homegrown encoding   
   was broken and it was easy to generate and check base64 (another   
   encoding/decoding program already existed).   
      
      
   Certainly _my_ rules would allow any sort of encoding of the payload,   
   but limited obviously by which chars were allowed. (For example I   
   don't like the "backtick" char (ASCII 0x60) and the backslash would   
   probably wreck havoc or be corrupted by transfer so I made sure that   
   those weren't used. Going so far as to move around blocks of code so   
   they disappeared, as they usually was being generated by offsets in   
   the second stage decoder.)   
      
   I think my allowed ASCII ranges could be summarized as   
   0x21-0x5b+0x5d-0x5f+0x61-0x7e.   
      
      
   --   
   MartinS   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)