Friday, October 26, 2007

Excel ate my election!

Excel ate my election! seminar
I'm currently on tour (!) in Scotland, visiting colleagues at Glasgow, St Andrews and Edinburgh universities.

Yesterday I did a seminar at St Andrews on the Open Rights Group's work observing e-voting and e-counting during the May elections. With impeccable timing, the Electoral Commission's independent Scottish Election Review was published this week, allowing me to update my slides.

ORG e-voting coordinator Jason Kitcat has already written about the review. But reading the full report gives an even more jaw-dropping picture of the fiasco that resulted from grossly inadequate planning, design and testing of ballot papers and e-counting systems in Scotland.

I am appalled that the Ministry of Justice is apparently still considering further e-voting trials in the UK.

UPDATE: Interesting response to the review from Secretary of State Des Browne:

"[T]here will be no necessity for electronic counting in elections, either for this Parliament or for the Scottish Parliament."

UPDATE 2: The Scotsman reports (via Open Rights Group): "CITY council bosses have secured more than £100,000 in compensation from the company behind the electronic count at this year's elections.

"The agreement comes almost six months after the fiasco that saw postal votes failing to arrive on time and election results delayed for hours because of problems with the counting software."


William said...

> I am appalled that the Ministry of Justice is apparently still considering further e-voting trials in the UK.

Are we *sure* they are? Is there evidence?

Ben said...

Whilst I would agree with your comment about poor planning and design I am would like you to expand on your rather vague comment about electronic counting.

My reading of the Commission report was obviously somewhat different to yours, as the Commission appeared to actually endorse the use of the technology for STV counting.

But why let the facts get in the way of a controversial rant eh?

Ian Brown said...

I would not characterise the following comment from p.106 of the Gould report as endorsing the e-counting systems used. Nor, judging by the press coverage of the report, would I characterise my post as a controversial rant. Mr Pot, meet Mr Kettle?

"In discussion with the various stakeholders involved in the electoral process, it became abundantly clear that the problems outlined above contributed to a loss of confidence in the electronic counting system amongst many of them."

Ben said...


We can go back and forth quoting from the Electoral Commission report all day. On page 119...

"The electronic count and procedures related to it came under particular scrutiny following the close of polls on 3 May 2007. We have found that many
of the problems identified, as well as earlier incidents where it seemed that technology was the driving force behind electoral procedures (rather than the
reverse), were directly attributable to legislative delays. Such delays would
ultimately influence the ballot paper design and thus the count, given the need for ballot papers to be compatible with the electronic scanners.

If the STV electoral system is here to stay, the electronic count cannot be reasonably abandoned. While there were some problems with the electronic count on the night of 3 May 2007, such as the database malfunction specific
to the DRS system that occurred at some counting centres, there is little doubt that the electronic count facilitates the counting of STV ballot papers and, in this respect, is preferable to a manual count. It is important to note
that our assessment has not found any evidence that the electronic count contributed to the number of rejected ballot papers.

All this considered, we would RECOMMEND that electronic counting continue for future combined parliamentary and local government elections or local government elections alone when the STV system is being used."

There were fundamental problems with the way the Scottish election was planned and implemented and these lessons need to be learnt.

There was a loss of public confidence in the counting system. The media coverage of the event guaranteed that and blogs such as this certainly aren't doing anything to help this perception.

The technology worked, the count was accurate and delivered within the agreed timescales. These are the facts.

Ian Brown said...

We could indeed quote all day from the report. On your assertion that "the technology worked" for example:

"A database malfunction within the DRS electronic counting system occurred at some count centres, and the impact varied from centre to centre. In the most extreme cases, occurring at five count centres, all counting activities had been completed but it was not possible to produce the validated results and therefore the completion of the count had to be suspended until later on Friday." (p.101)

"It was not until 12.30am on Friday that the problems appeared at the Edinburgh count... A potential solution was identified between 3.00 and 4.00am... The solution was then tested and implemented fully on the Edinburgh servers, which were operational by about 5.00am. By that time, a number of counts had already been suspended until later on Friday morning. Further testing was subsequently carried out in Edinburgh to ensure that the solution was suitable for all count centres. Senior members of the technical team travelled to the West Lothian count to undertake further testing and a patch was created and published via the DRS intranet at approximately 8.30am on Friday. This patch could subsequently be downloaded by DRS staff at each count centre and applied to the system." (p.101)

I don't know about you, but I prefer that mission-critical software isn't patched in the middle of the night and that such patches are not further applied across a system without more than a few hours of testing - especially given that "A subsequent investigation undertaken by DRS following the count determined that while the fix identified on the night had resolved the issues, the team had not identified the root cause of the problem." (p.102)

Ben said...


Clearly there were issues and the report highlights these. However if the technology hadn't worked then the count would of been called into question and you would have yourself a real story.

The Electoral Commission have satisfied themselves of that fact and have recommended the on-going use of electronic counting but perhaps you know better than them.

If you are truly concerned about the Scottish election then perhaps your time would be better focused on looking into the electoral procedures and legislative delays that caused the majority of the problems, but I guess that just ain't as interesting...