CA> Archived software is compressed on a sliding-Huffman code
 CA> that changes depending on the frequency of byte codes in ALL
 CA> the files compressed.  This means the same virus would appear
 CA> as different bytes in EVERY archive it was in.

true...

 CA>  How could any
 CA> software accomodate this and 'recognize' the virus?  In spite
 CA> of this, they 'scan' archives for virus - NOT!

quite simple, actually... they scan it as they unarchive it... they don't
have to unarchive to the disk, either... they can unarchive into a
"sliding buffer"... i can definitely tell when they start
scanning thru my filebase when i forget to tell them not to...

)\/(ark