Up: Home page Previous: Issue 21

Voting matters - for the technical issues of STV

Issue 22, July 2006

(Complete issue in PDF, 204Kb.)

Editorial

There are five papers in this issue, the first three being:

Jonathan Lundell: Random tie-breaking in STV.
Although Voting matters has had several papers about tie-breaking, one can see that there is still more to be said on the matter.
David Hill: Implementing STV by Meek's method.
David Hill has provided an implementation of Meek's method for many years. This implementation has been taken as the `definition' of the method for the New Zealand elections. In this paper, the details of this implementation are described and contrasted with that of the original Computer Journal article.
Robert Newland: Computerisation of STV counts.
Although Robert Newland died in August 1990, readers may well be surprised at the relevance of this paper for today. Up to his death, he was the leading technical expert on STV within ERS. This paper was located by David Hill and since it was never published, printing it here seemed appropriate. It is hoped that readers will respond to the suggestions made.

The final two papers have a common theme: the form of STV proposed by British Columbia and now being considered for the Scottish local elections to be held next year.

Jeff O'Neill: Comments on the STV Rules Proposed by British Columbia.
This paper presents the details of an implementation of the British Columbia rules which has been available on the Internet for some time. It is a very simple version of STV in computer terms. Several issues arose from this work which are detailed in the paper.
James Gilmour: Developing STV Rules for manual counting to give effect to the Weighted Inclusive Gregory Method of transferring surpluses, with candidates' votes recorded as integer values.
The paper is a complete contrast to the previous one. Like the previous paper, the aim is to transfer surpluses by considering all papers, not just the last batch that gave rise to the surplus. The contrast is in its presentation as a manual counting process and the provision of the conventional result sheet. One novelty is (at least within the UK) that the calculations are undertaken with high precision, but the results are presented as integers.
James Gilmour has produced a proposal and sent it to the Scottish Executive. This proposal, slightly modified, is now on the McDougall web site. Hence the article provides the rationale and background to the proposal.

It is hoped that the contrast between the two methods above will clarify the choices to be made for the Scottish elections. The final choice will be awaited with interest.

Two other items may be of interest to readers. Firstly, the final report on electronic voting in Ireland is due out shortly and will be found at: http://www.cev.ie/. Secondly, it has come to my attention that the British Computer Society elect their council by STV but do not provide a result sheet to their electorate - only a list of those elected. Since the transfer of votes will not be visible, this seems to me to be STV in name only. Do readers have other examples of this?

Readers are reminded that views expressed in Voting matters by contributors do not necessarily reflect those of the McDougall Trust or its trustees.

Papers with citations

Jonathan Lundell: Random tie-breaking in STV. (p1-6, PDF 48Kb)
David Hill: Implementing STV by Meek's method. (p7-10, PDF 36Kb)
Robert Newland: Computerisation of STV counts. (p11-13, PDF 28Kb) [23, 1-2, PDF]. [23, 3-9, PDF].
Jeff O'Neill: Comments on the STV Rules Proposed by British Columbia. (p14-20, PDF 60Kb)
James Gilmour: Developing STV Rules for manual counting to give effect to the Weighted Inclusive Gregory Method of transferring surpluses, with candidates' votes recorded as integer values. (p21-25, PDF 40Kb) Since publication an error has been noted in this article for which an author's correction is available here. A formal correction will be published in the next issue. [23, 3-9, PDF].

Main index Author index