Re: Re: Election entropy as CGI interface

From: David Mertz <voting-project_at_gnosis_dot_cx>
Date: Fri May 07 2004 - 22:20:59 CDT

On May 7, 2004, at 10:38 PM, Steve Chessin wrote:
> Thanks! I did find the "instructions" a bit terse.

Well, yeah.

> Is "multi 1 10" the same as "single 10"?

Actually, it is. Or rather, "multi 1 10" might be thought of as a
degenerate case of multiple-selection.

I considered only providing the multi option, but since
single-selection is the most common style for US ballots (by a good
margin), I felt it would be easier for user to give it its own type.
Not maximally orthogonal in the data file... but more readable.

> I also don't understand what "alphabet size" means. It seems to only
> affect the self-delimited result. (I gave it this example:

Right. According to Arthur (I haven't read the details of why), the
optimal size for self-delimiting values depends on the size of the
symbology being encoded to. Some barcodes have 16 different codes,
some have 100, some have 8, etc. The best self-delimited encoding
depends on the barcode chosen. But the globally optimal encoding does
not depend on this (we just need a bits2baseN() function for the
conversion).

> Alphabet sizes of 1 and 3 worked, but 2 gave me a divide-by-zero
> error.)

Well... OK, don't do that then :-).

A symbology with two different letters is rather degenerate--it's known
as binary, actually. A symbology with only one letter seems even worse
though! I'm sure I could invent some descriptive message for those
cases, but basically, they're not interesting cases to worry about (so
if my results are wrong/meaningless, who cares).
==================================================================
= The content of this message, with the exception of any external
= quotations under fair use, are released to the Public Domain
==================================================================
Received on Mon May 31 23:17:25 2004

This archive was generated by hypermail 2.1.8 : Mon May 31 2004 - 23:18:16 CDT