Scytl company PNYX approach compared to TLV

["David Webber \(XML\)" <david@drrw.info>]

I found this link off the NIST vote.nist.gov site.

http://www.scytl.com/docs/pub/a/PNYX.DRE-WP.pdf

Essentially these folks have implemented the MIT "frog" approach.

This comes closest to the TLV approach - because it does provide
a means for dual verification.  However - there are serious flaws in
their approach.

First - while their "verification unit" (aka MIT 'frog') does an electronic
verification - its just fancy flashing lights - compared to a real physical
paper ballot print-out and direct human verification of that.

Their argument that the paper can be tainted fails when you use 
OASIS EML to ensure only required information is passed between
the voting unit and the verification unit.

Again - this is a significant flaw in their design - they are not using
XML formats that can be verified.  Therefore their verification step
is in question - since it too can fake out the voter (show them one
thing - record another).

Anyway - useful to know that people are at least trying to do it 
right!   

Maybe someone can suggest to PNYX that they upgrade their
system to support EML 4.0 and direct voter verification?!

DW