Best article vote tally - WE HAVE TWO WINNERS!

[Steven Schveighoffer]

On Thu, 09 Jun 2011 15:35:44 -0400, Robert Clipsham  
<robert at octarineparrot.com> wrote:

> On 09/06/2011 20:21, Steven Schveighoffer wrote:
>> On Thu, 09 Jun 2011 15:02:08 -0400, Walter Bright
>> <newshound2 at digitalmars.com> wrote:
>>
>>> On 6/9/2011 11:03 AM, Robert Clipsham wrote:
>>>> So there is going to be a next one?
>>>
>>> Yes, maybe in 6 months or so. I'm very happy with how this one turned
>>> out.
>>>
>>> But next time we need to devise a tie-breaking rule. Any suggestions?
>>> A runoff?
>>
>> We're all developers here, I think people might be open to an instant
>> runoff:
>>
>> http://en.wikipedia.org/wiki/Instant-runoff_voting
>>
>> Essentially, you rank the articles 1 to x, and then the algorithm
>> figures out the winner. It's still possible to have a tie, but unlikely.
>
> I thought about mentioning this, but decided against it. I believe FPTP  
> is the best way to vote for this kind of competition.

I wonder if there's some form of instant runoff that only breaks ties.   
That is, a runoff only occurs between ties, with the people who did not  
vote for the tied candidates getting their secondary votes counted.  I  
suppose most voting systems are for votes that count in the hundreds of  
thousands to millions, so there is very little chance of a tie.  So this  
might be unexplored territory...

I just am not keen on the idea that someone can vote for candidate A, then  
when A ties with B, vote for candidate B in the runoff.  Instant runoff  
appeals to me because you have to write down your preferences up front.

>
>> I think for the next time, someone should write a newsgroup-to-vote
>> program that automatically counts the votes (must be in D of course!)
>
> Of course, then that one person who doesn't format their vote quite  
> right loses out...

Well, we can make it simple:

Mark your preference in this box (1-5):
  |
  v
[ ]  article 1
[ ]  article 2
...

Another (really good) option is to use a web-based voting system, which  
makes things verifiable.  You still need something to verify the user has  
posted to the NG in the past.

>>> BTW, there's nothing in the rules preventing an author from tooting
>>> his own horn and doing a bit of marketing of their article(s) for  
>>> votes!
>>
>> We're developers, not politicians :) If you allow this, then we'll have
>> to start creating youtube ads showing the other articles' past records
>> of infidelity and such, and it just turns ugly.
>
> Given the rule that voters must have used their handle here before,  
> that's not going to happen. I won't comment further to avoid ranting.

I hope you didn't think I was serious, though I don't see how it could be  
seen that way.  If I offended, I'm sorry.

-Steve