Is pro-VVAT in the same category as anti-MMR?

williamgrogan

Victor, I agree with nearly everything you say in your last post, except your conclusion.

Your opening comments would indicate that you were going to present evidence that eVoting is unreliable. You didn’t.

I too worked in IT in construction and I agree that the fraud there is awesome.

The EDI company that was tasked with spreading the use of electronic data interchange didn’t have an email address until relatively recently either. I wonder is that why EDI is almost non existent?

Demanding VVAT with eVoting is the same as doing your valuations with the laptop and then repeating them manually with the old “dim” books (Is that what they were called?) just to make sure the programs are OK for evermore?

I’m still waiting on anyone to support the statements of the two TD’s. Do you agree that, “they were economical with the truth”, to use a well known euphemism? It’s not just these two statements but that these are proof of the point I am making that the anti-eVoting lobby are lying by omission and exaggeration just like the anti-MMR people. (Not to mention a certain charity that says thousands of Russian children need de-toxing in Ireland.)

williamgrogan

see how the slanted stories spread.......

http://www.dmeurope.com/default.asp?ArticleID=1541

Is this accurate?

However, a leaked report commissioned by the Irish government uncovered that the Nedap voting machines do not produce an “audit trail”.

Look at the meaning of the word "had" below. Is this true? It sounds like eVoting is now under threat.

The Irish government had planned to deploy these devices (also used by the Dutch) in elections for the European Parliament in June.

Was the information obtained under the freedome of information act, leaked?

However, a leaked report..

When the Luddites win....

http://www.tynetees.tv/_lab/news/news_headlines.asp?newsitemid=6505&type=news

williamgrogan

I didn't know there was a roadshow to demonstrate the new and highly accurate eVoting machines.

see here for towns & dates

http://www.electronicvoting.ie/english/roadshow.html

RainyDay

No reaction to today's events, William - Cat got your tongue?

davros

Clearly the Independent Commission waited for wg to go on holidays before releasing its report

williamgrogan

Nice of you to miss me.

I is on holidays in Memphis, TN. Big Blues event.

I have read the report and will comment. It sounds like they couldn't think of a good reason to shelve it but shelved it anyway just to be on the safe side. Hopefully they won't set up a commission to investigate the reliability of air travel or it will be suspended using the same logic.

PS

I too am waiting on a reply to my question, whether the comments by the two TDs were in effect lies.

ShaneHogan

Originally posted by williamgrogan
I have read the report and will comment. It sounds like they couldn't think of a good reason to shelve it but shelved it anyway just to be on the safe side.

Congratulations, William - this is just your best one yet. One can only assume that the hot sun in Memphis has either gone to your head or is glaring so much on your screen that you couldn't actually read the report. Either that or you display an amazing ability to ignore the facts which you don't like - so much for the importance of skepticism and looking at the hard data.

Just for the record, in case you missed the many, many reasons why this system cannot be implemented,

- The so-called 'security hardened' PC's proved quite easy to break into.
- There has been insufficient integrated end-to-end testing, and none of this testing has been done independently.
- The commission found one bug in the count software (and they had let than two months to review it).
- The actual version of the software intended to be used for the June elections was still not available to the commission for review/testing, so no-one (not you, not me, not Nedap and not Dept of Environment) can speak with any certainty about the reliability of this version.

Looks like it is time to retire gracefully on this one, William....

williamgrogan

I am just back and will reply in detail after I catch up with my work. However, ………

The so-called 'security hardened' PC's proved quite easy to break into.

as easy as a ballot box? Most of the security is human related and is at least as good as the existing security associated with the manual ballot box.

There has been insufficient integrated end-to-end testing, and none of this testing has been done independently.

The Space Shuttle was first launched without end to end testing. End to end testing is just more testing. If the sub systems were tested then all end to end testing does is more testing of the interfaces and a further improvement in reliability from say 99.990% to 99.991%. It is you lot brought up this point. That doesn’t mean that it’s essential. End to end testing was carried out in the “dry run” in the last elections where everyone using the system expressed satisfaction with it. The commission did stress that the software passed all tests and that they had confidence it did record and count the votes accurately.

The commission found one bug in the count software (and they had let than two months to review it).

Minor as I understand it. Do you know what this bug was exactly?

A point that has not been made up to now is that you lot have been saying that from a statistics point of view the software with 200,000 lines of code must contain bugs and I agree. There is however a major point that has not been made. Software can contain thousands of bugs that will never actually cause a problem or if they do the problem will not effect the outcome or if it does it will only effect the outcome in a trivial way or a “cosmetic” bug. Many bugs are omissions of tests for events that will never occur, e.g. maybe there is a bug that means if there are 32767 candidates that the 32767th will not have his vote counted. It is a bug but as there is never going to be 32767 candidates then it will never matter. I think this is a very important point because we are only concerned with bugs that affect the outcome.

The actual version of the software intended to be used for the June elections was still not available to the commission for review/testing, so no-one (not you, not me, not Nedap and not Dept of Environment) can speak with any certainty about the reliability of this version.

It is not necessary or practical to do a total system test after a minor version change in any system, as I am sure you well know.

Finally I am totally opposed to VVAT and my first reading of the report, which I did point out was not a careful study, does not indicate support for VVAT despite overwhelming support from those making the submissions.

ecksor

Originally posted by williamgrogan
Finally I am totally opposed to VVAT [..]

I thought you were opposed to paper.

williamgrogan

I've a foot of f*****g paper on my desk after my trip not to mention 1000 emails. Go away.........



PS

Yes, I have no problem with electronic audit trails I have one in my system.

ecksor

Originally posted by williamgrogan
Yes, I have no problem with electronic audit trails [..]

And you've even said that you'd have sympathy if the campaigners for VVAT were looking for an electronic VVAT, rather than a paper one. Can you provide an example of a non-paper based VVAT that has stood up to scrutiny?

ShaneHogan

Originally posted by williamgrogan
Yes, I have no problem with electronic audit trails I have one in my system.

Any electronic audit trail for an eVoting system is not worth the paper its not written on. Any attack on the vote data simply has to also attack the audit trail.

as easy as a ballot box? Most of the security is human related and is at least as good as the existing security associated with the manual ballot box.

You miss the point. Any attack on a ballot box would impact maybe 1% of the votes of a constituency. It would also be effectively impossible to disguise an attack on a ballot box. If you wanted to 'stuff ballots' or replace the contents of the ballot box, you would need to either know exactly how many votes went into the box originally or have sufficient time to count the papers before replacing them. The chances of doing this undetected given that ballot boxes are under Garda supervision and are frequently checked by the election agents of the candidates and are in the public view once they arrive at the count centre.

By contrast, anyone who gets access to the count centre PC has the entire vote database at their discretion. It would be a relatively straightforward matter to load a piece of attacking software (maybe by popping a CD in the drive and using autorun to activate it - don't even need to touch the keyboard – or by connecting a pen drive to the USB port) which simply updates the vote details in the Access database to ensure that one candidate/party wins. Such an attack (if carried out effectively) would leave no trace – no-one would know that the attack had happened. Hence the huge danger of having an insecure PC controlling the count.

The Space Shuttle was first launched without end to end testing.

Actually, you’ve just sparked a memory of John Craven’s Newsround showing the Shuttle being carried up to height on the back of a specially adapted Jumbo Jet and then ‘set free’ to test its landing capability – Would that count as an end-to-end test? Well, I don’t really know enough about Space Shuttles to comment. I do know that comparisons to Space Shuttles and banking systems are just not relevant.

End to end testing is just more testing. If the sub systems were tested then all end to end testing does is more testing of the interfaces and a further improvement in reliability from say 99.990% to 99.991%.

The example figures that you quote are meaningless. No piece of software developed in this way will ever be 100% accurate, so you will just never know when you are 90% done, or 99% done or 99.99% done. End-to-end testing does a lot more than testing of the interfaces. It tests how the entire holistic system is going to work in the real-life environment. For example, the ‘randomisation’ feature of the IES count software was disabled for all of the unit tests, so no-one really knows ho this works in real life. Proper end-to-end testing would clearly demonstrate the success or failure of this feature. End-to-end testing also serves to test the manual control processes which humans operate around the system For example, the difficulties with the paper forms used by election staff which went unnoticed until the 2002 trails would almost certainly have been picked up by any half-decent end-to-end test. It is patently obvious to any unbiased IT professional that the level of end-to-end testing carried out by the Dept (i.e. run two local authority ward counts through the system, and ignore the differences between the test results and the original manual counts) was grossly inadequate.

It is you lot brought up this point. That doesn’t mean that it’s essential.

Well, it doesn’t really matter what you think or what I think. What really matters is what the CEV thinks – and they clearly thought it important enough to highlight in their report.

End to end testing was carried out in the “dry run” in the last elections where everyone using the system expressed satisfaction with it. The commission did stress that the software passed all tests and that they had confidence it did record and count the votes accurately.

The trial run was a trial run, not a test. No-one knows what votes were entered into the system on that day, so no-one can speak with any authority as to how accurately the votes were counted. We can say that the results were broadly in line with expectations and with the manually-counted constituencies. But ‘broadly in line’ is just not good enough, given the importance of this system to our democracy. No commercial IT system would be installed on this basis (i.e. we didn’t really test all the bits together, but when we ran the first month’s accounts, they were generally in line with what we saw last year). I’d be fired if I attempted to go live with a system on that basis.

There is however a major point that has not been made. Software can contain thousands of bugs that will never actually cause a problem or if they do the problem will not effect the outcome or if it does it will only effect the outcome in a trivial way or a “cosmetic” bug.

It is possible that you are right here. It is also possible that you are wrong here, and that the remaining bugs in the system WILL cause problems and WILL effect the result. No-one knows for sure. What is clear is that the Dept & Nedap have been operating in an ‘iterative’ model, i.e. Nedap fix a few more bugs, Dept & ERS retest, Nedap fix a few more bugs, Dept & ERS retest. As a result, it is clear that they will just never finish – the software will never be 100% complete. There is no sound basis for implementing such a system in this environment.

It is not necessary or practical to do a total system test after a minor version change in any system, as I am sure you well know.

What minor version changes? How have you concluded that the bug fixes they are implementing are minor version changes? Unless you have specific inside information that has not been published elsewhere, you have no basis for this conclusion. Given the fundamental importance of this system, there is no excuse for issuing ‘minor version changes’ two months before the expected go-live date.

my first reading of the report, which I did point out was not a careful study, does not indicate support for VVAT despite overwhelming support from those making the submissions.

The CEV report does not go all out and state that VVAT is mandatory. However, they clearly considered VVAT important enough to be worthy of mention – despite that fact that it falls outside the narrow terms of reference which Cullen gave them. That speaks volumes. It also states that the absence of VVAT significantly raises the bar on the level of testing required. The Dept failed to leap over that particular bar.

You can try & do a King Canute and tell the tide to stop coming in, William – but it just hasn’t worked. The CEV tide has washed away you, Martin Cullen and the guys & gals in the Dept of the Environment.

Regards - Shane

PS Can I ask that in the interests of maintaining a civil debate, you refrain from referring to anyone or any group as ‘you lot’. It is just a bit impolite.

williamgrogan

What with jet lag and a mountain of paper I still don't have time to reply fully but I will.

However now that we have a "surprise" result in India after 700,000,000 voted using eVoting (and I suspect using a relatively primitive eVoting system) how long will it be before this "surprise" result is quoted by a TD in the Dail as evidence of the "many failures of eVoting"?

Victor

Originally posted by williamgrogan
However now that we have a "surprise" result in India after 700,000,000

O_o they were saying only 380m voted earlier.

williamgrogan

I stand corrected, there was in fact 670,000 electors.

I have since listened to RTE's 17:00 news where their reporter in India said the eVoting was a big success and commented that it was odd that a country of a billion people could introduce eVoting and our country with 4 million couldn't.

Apparently DRE's were dropped by parachute into some areas, how's that for end to end testing? Instruction leaflets on what to do were also dropped by plane.

Then again India is a leading software development centre so maybe there are fewer Luddites and old codgers there than here.

RainyDay

Any chance you could give the meaningless broad-brush generalisations a rest, William. There is no comparison between the Irish & the Indian systems. For a start, the Indians paid out about €100 per voting maching, by comparison to the €4k (IIRC) which Minister Cullen was happy to hand over. As Pat Rabbitte said, it's not the indians you have to worry about, it's the cowboys from FF & the Dept of Environment you need to keep an eye on.

williamgrogan

I think calling FF & the Dept of Environment cowboys is a broad generalisation. Don't you?

I think we have one of the most honest civil servants in the world.

So perhaps you could give the broad-brush generalisations a rest.

I had already said "...and I suspect using a relatively primitive eVoting system". Anyway you can't compare prices between India and here, e.g. a software engineer in India gets a fraction of what one is paid here.

RainyDay

Fair cop on the generalisations, William. Can we leave India to one side and focus back a little closer to home.

I was referring specifically to the hardware cost of the machines, not the supporting software. Any good reason why we needed to buy the Rolls Royce of voting machines?

Victor

Originally posted by williamgrogan
I think we have one of the most honest civil servants in the world.

Which one is the honest one?

williamgrogan

I notice that there is almost no mention of any anti-eVoting lobby or pro-VVAT argument in India. That's another similarity with the anti-MMR campaign. The anti-MMR lobby simply doesn't exist in many countries and then in others its big news. The uptake of MMR is very low in Ireland and the UK but very high in most other EU countries.

......... re the Anti-MMR Campaign

Well that's one out of the way!

Consumer Health Digest #04-21
Your Weekly Update of News and Reviews
May 26, 2004
Current # of subscribers: 9,103

IOM debunks alleged vaccine-autism link.

An Institute of Medicine (IOM) expert committee
has concluded that neither thimerosal nor the
measles-mumps-rubella (MMR) vaccine are
associated with autism and that the hypotheses
connecting them are not worth further study. The
committee report updates two IOM reports
published in 2001 that found no association but
recommended further research. Since that time,
five large epidemiologic studies conducted in the
United States, the United Kingdom, Denmark, and
Sweden have found no association between
thimerosal-containing vaccines and autism, and 14
large epidemiologic studies consistently showed
no association between the MMR vaccine and
autism.

I wonder will the UK Independent apologise. How many children will get sick and even become disabled because of its anti-MMR campaign?

dogs

Originally posted by williamgrogan
I notice that there is almost no mention of any anti-eVoting lobby or pro-VVAT argument in India.

You noticed did you ? You're on a lot of Indian tech-lists then ? You swung by on a recent jaunt from your adopted French village to Roches Stores, Cork ?

When you're done with the unfounded, unreferenced statements try google. It's been discussed on mailing lists there, this article mentions at least one of the objectors.

There's not much point me re-stating the technical facts of this. It hasn't worked yet. Let me try some other facts.

• You are a skeptic.
• You participate in democratic elections.
• These elections grant power and privilidge to a select few in our society.
• You grant implicit trust that these people will not abuse that power or privilidge specifically in the mechanism whereby they are elected.
• You grant implicit trust that a system without comprehensive end-to-end testing works as specificied, without error due to technical failure, human error or malicious intent.
• You are a skeptic ?

williamgrogan

Dogs reply to my post does not address the point I made. Instead he tries to slag me off in some way related to personal information I supply. I might add again that I use my name and not the name of another species. (Incidentally I got a reply from Roches Stores & the supplier of the “magnetic reflexology insoles”) and when I have the time will amuse you with it.)

The anti-MMR movement was only active in some countries. In others it simply was not an issue. To me this is evidence of the irrelevance of both “anti” arguments. The same seems to be the case with anti-eVoting. It is a big issue in the USA and Ireland and maybe Australia but nowhere else that I can see. It also seems that in Ireland we picked up the “virus/meme” from the English speaking USA. All this goes back to a few academics in America.

An RTE reporter specifically said that the eVoting had been a big success in India and pointed out that there was an overall acceptance of it. My Google newsalerts has delivered many articles from Indian newspapers and none of them indicates any anti-eVoting movement.

The same system as used here is in use in other EU countries and little or no anti-eVoting seems evident there either. The same incidentally can be said of Nuclear Power. France generates about 75% of its electricity by NP and after living for 6 years in France I saw no evidence of an anti-NP movement. That’s not to say there is none at all but it’s simply not an issue. Here, there are people that would die to stop it.

Its all an illogical fad.

RainyDay

Hi William - Rather than drawing broad-brush conclusions from such lightweight sources as "RTE reporter specifically said that the eVoting had been a big success in India and pointed out that there was an overall acceptance of it", why don't you go back and answer the detailed issues raised earlier, as you promised to do in this post.

williamgrogan

I will, it will take several hours to study the report and reply to it. I don't have several hours to spare at the moment. It is astonishing that as far as I can see no one else including the Government have made a substantial reply to the report or an analysis of it.

How is an RTE reporter who was in India to cover the election and who travelled around monitoring it a "lightweight source"? I would think he was a heavyweight source.

You don't address my point. In India and other EU countries the anti-eVoting lobby is almost non existent. This anti-eVoting “meme” seems almost exclusive to the USA and Ireland. The anti-MMR campaign was the same. So was the anti-GMO campaign. No problem in the USA, big problem in Britain.

Remember, I believe that the over the top reaction to and exaggerated fear of eVoting is Luddism, as is the anti-GMO, the anti-NP, anti-fluoridation and the anti-MMR movements. All share many common traits including an inability to access the actual risk and the consequences of failure. I have just ordered a book about “Great Battles in Technology”, which may throw some light on this issue. The author says that, “the success of technology is related to the determination of its proponents” (and presumably also to the determination of its opponents).

ecksor

As well as not addressing the apparent inconsistencies that I pointed out in my last post on this thread I have to draw attention to:

All share many common traits including an inability to access the actual risk ...

You didn't pursue your risk analysis earlier either once I commented on what I perceived were the shortcomings in your first attempt.

williamgrogan

Some more "lightweight" analysis of the Indian election.

"In praise of electronic voting" from an editorial by The Hindu, an Indian daily

http://www.abs-cbnnews.com/NewsStory.aspx?section=Opinion&OID=51940

dogs

Originally posted by williamgrogan
I might add again that I use my name and not the name of another species.

I'm sure there's a Gathering card that covers this...

(Incidentally I got a reply from Roches Stores & the supplier of the “magnetic reflexology insoles”) and when I have the time will amuse you with it.)

Excellent, that is good, there's far too much sympathetic press about magnetic cures already ....and yet people also complain about the risk of cancer from electromagnetic fields from powerlines, microwaves, etc. I'm getting off topic but do post the letter sometime.

The same system as used here is in use in other EU countries

But it's not y'see. It's not the same system by a long way. Australia even went as far as publishing in their code, we've had to make do with two reviews. Government appointed reviews. There would be no bias in selecting who to review the project after sinking in excess of 40 million euros ?

williamgrogan

Interesting New York Time article

If election officials want to convince voters that electronic voting can be trusted, they should be willing to make it at least as secure as slot machines. To appreciate how poor the oversight on voting systems is, it's useful to look at the way Nevada systematically ensures that electronic gambling machines in Las Vegas operate honestly and accurately. Electronic voting, by comparison, is rife with lax procedures, security risks and conflicts of interest.

On a trip last week to the Nevada Gaming Control Board laboratory, in a state office building off the Las Vegas Strip, we found testing and enforcement mechanisms that go far beyond what is required for electronic voting. Among the ways gamblers are more protected than voters .......................

The full article is here

williamgrogan

An interesting article on the negative side of the Precautionery Principle in Spiked.

some extracts

'If everything we did had to be absolutely safe, risk-free, proven to have no adverse outcomes for anyone or anything, we'd never get anywhere. Buildings wouldn't go up, planes wouldn't get off the ground, medical breakthrough would come to a standstill, science would be stifled…. Shall I go on?'

{WG - I might add ... and eVoting}

....

Professor Sir Colin Berry is not a big fan of the 'precautionary principle', the idea that scientists, medical researchers, technologists and just about everybody else these days should err on the side of caution lest they cause harm to human health or the environment. Berry is one of Britain's leading scientists;

....

'Almost no new technology can be assured to be risk-free. If your position is that you don't accept any incremental risk, you are in effect saying no to all new technologies, whether it be a better anaesthetic, a better car, a better aeroplane, a safer environment for children - in fact anything worth having.'

complete article

The reason I bring this up is that it is another similarity between the anti-MMR debate and anti-eVoting.

I have already said that there we can expect some small problems with eVoting but that they will be resolved and the system would evolve. The commission even said they believed that the system would work and also said that deciding to shelve it was easier than proceeding. This is the Precautionary Principle at its negative worst.

A similar argument was made with MMR, because Wakefield claimed their was a small chance of a link between MMR and Autism he suggested stopping the MMR.

PS

I met a radiographer at a function last night who has an Autistic child and she suspected that it might have been caused by MMR. I reassured her it wasn’t. What’s amazing is that radiographer’s are scientifically trained, they are supposed to have a strong Physics background!

I just spotted another related article.

Science, risk and the price of precaution

In the run-up to spiked's conference Panic Attack: Interrogating our obsession with risk, taking place at London's Royal Institution on Friday 9 May, we asked 40 members of the international scientific community to list what significant discoveries and achievements would have been limited or prevented, if science at the time had been governed by the precautionary principle that dominates science today.


Between them, respondents came up with an A-Z of historic achievements that would have been thwarted by the precautionary principle:

here

RainyDay

Originally posted by williamgrogan
The commission even said they believed that the system would work

Please confirm exactly what section/comment from the CEV report you are referring to here?