Should we discontinue EVMs?

[Tanaji]

The other thing is no amount of checking is enough if you are a conspiracy theorist. For example Rahul Mehta uvacha:

And audits are useless. I know how good PwC auditors were when they "audited" Satya accounts. So I dont trust audits.

So even if you do audit something, there are always characters that will pop up and say I dont believe this because it contradicts my belief. In fact Rahul Mehta is on record in another thread where when proved wrong by impartial, impeccable references, he stated that since the references did not agree with his assumptions/beliefs, they were wrong and I needed to provide statistics that agreed with him.

In face of an attitude that is devoid of logic and conveniently bends space and time laws to suit them, it is impossible to argue or prove anything.

[Sanku]

All is lost run run flee flee, it is the end of civilization. Return to the old method of substience agriculture and keep a stocked armory in your basement.

Practice it often

Form vigilante bands of people you know by blood.

Always have a collective decision by them signed in blood before doing anything.

Woe is me... woe is me....

[Rahul Mehta]

So, the ballot-box stuffing may be gone, but IMHO there remains a clear and present danger of industrial scale rigging.

The ballot stuffing was 1% reality and 99% an urban legend.

First, the local presiding officer has to be bribed. Not everyone would co-operate. Next, there are 2-3 more Govt employees in booth, and all of them have to bribed.

Second, whatever may have been that x% reality in ballot stuffing, one can reduce that putting cameras in booths that take pix every 30 seconds and put it on central server using GPRS modem. So candidates, media and anyone can see any booth live for a fee to cover costs.

[Sanku]

For the record, I agree with a lot said about Mr Singh on the forum. I see no +ve with him at helm and am no fan of congress either (clearly seen by my prior posts on election thread)

I also agree with Pranav when he raises a lot of potential holes in the system -- with respect to Pranav's list though I dont think it can be stitched together neatly enough to do rigging, this however does not reduce the importance of his list and various points when considered separately per se.

As far RMjee "logic" is concerned that since Shri MMS is bad for India, CIA has rigged the EVMs, is like saying since I dont like Mangoes some one has gone around making sure that Orange trees now produce mangoes.

[Tanaji]

So, the ballot-box stuffing may be gone, but IMHO there remains a clear and present danger of industrial scale rigging.

The ballot stuffing was 1% reality and 99% an urban legend.

First, the local presiding officer has to be bribed. Not everyone would co-operate. Next, there are 2-3 more Govt employees in booth, and all of them have to bribed.

Second, whatever may have been that x% reality in ballot stuffing, one can reduce that putting cameras in booths that take pix every 30 seconds and put it on central server using GPRS modem. So candidates, media and anyone can see any booth live for a fee to cover costs.

Ooooooooooooh GPRS modems!!

Have you audited the GPRS modem? Have you audited the central server? Have you audited the cell phone network? How do you know someone is not sending false pictures showing everything is good? It is TRIVIAL hacking cell phone networks...

[krishnan]

So what is the conclusion? EVM were rigged or not?

[Rahul Mehta]

I do not own this site. Nor did I post even one line on that site.

-----------------

http://government.wikia.com/wiki/EVM_Annexure_IV

The ability to create machine votes in a multitude of ways

1. About 200,000 EVMs were acquired in Jan. 2009 with modified programs (for date/time stamping); thus two types of EVMs were used, the old units without date/time stamping (about 11 lakhs of them) and the rest with the new 'improvised' feature claimed by BEL, the company which made the program change.

2. With this program change, an entry program is provided for any programmer to modify the PROM (Programmable Read Only memory) since time is determined outside of the balloting unit and thus beyond the control of the EVM system.

3. A selective rigging of these 200,000 EVMs would be adequate to impact the results in about 70 constituencies (say, those won in UP by Congress, those lost by opposition in Tamilnadu or Uttarakhand or Punjab, those won by BJD in Orissa). A remote control was possible to record, say, 4 votes for the desired party for every vote polled by any opposing party.

4. EC had also lost control over the EVM system because BEL, ECI contracted to make the systems had subcontracted the work to private contractors. No credible claims have been made by CEC about the systems audit conducted on the acquired EVMs and during the electoral process. CEC has clearly goofed up relying upon a scrappy Indiresan Committee report (just compare it with the professionalism with which system auditors audited the EVMs in USA).

5. India’s e-elections rigged? Saturday, July 04, 2009 NEW DELHI: The Indian Election Commission (EC) could be sitting on a major election-rigging scandal, following a presentation on Friday showing how the software used in the electronic voting machines (EVMs) can be manipulated. Omesh Saigal, an engineering graduate and former Delhi chief secretary, stunned the EC with a presentation showing that the software used in the EVMs can be manipulated to favour a particular party or candidate. Following the presentation, Chief Election Commissioner (CEC) Navin Chawla ordered an inquiry into the possibility of such rigging during the recently concluded elections in India and Indian-held Kashmir (IHK). Deputy Election Commissioner Balakrishnan was asked to conduct the inquiry on the basis of a report handed over by Saigal to the CEC, along with the software he had developed to show how the e-voting machines could be rigged. Saigal, who is an Indian Institute of Technology (IIT), New Delhi alumni, demanded an urgent check of the programme that runs the EVMs used in elections since 2004. The demonstration showed that after just keying in a certain code, the EVMs put every fifth vote in favour of a certain candidate. In his letter to the CEC, Saigal alleged that the EVM software had not been checked by the EC since the machines were manufactured more than 6 to 7 years ago. His argued that the EC merely relied on the certificates provided by the manufacturers, the government-run Bharat Electronics Limited (BEL) and Electronics Corporation of India Limited (ECIL). He alleged that the two firms had subcontracted private parties who actually provided the certificates. http://www.dailytimes.com.pk/default.as ... 2009_pg7_4, http://news.rediff.com/report/2009/jul/ ... rigged.htm

............................

------------------------

Tanaji,

Did you trust PwC a year ago? They were the most trustworthy audit organization in India. I did not trust them even then.

My point : show me a process/equipment that is physically less riggable than paper ballots. I dont trust audits as they are all no better than PwC. Now are you claiming that you would need more field agents to rig EVMs than paper? I see that you still have not written any post on COMPARING costs/security between EVM and paper. You have spent enough time in labeling me. Now pls spend some time in comparing COSTS of EVM vs paper.

[Sanku]

So what is the conclusion? EVM were rigged or not?

Depends on which universe in the multi verse continuum you are currently a part of -- as far as the widely accepted terrestrial reality exists -- It appears impossible.

But then we may be inside the matrix and I never know.

[Raju]

So what is the conclusion? EVM were rigged or not?

by all accounts the problem were in the later set of EVMs in my opinion. Until ECIL and BEL manufactured their own sets without subcontracting any of the work, it was all still relatively fair and transparent.

[Sanku]

by all accounts the problem were in the later set of EVMs in my opinion. Until ECIL and BEL manufactured their own sets without subcontracting any of the work, it was all still relatively fair and transparent.

At best what can be said is that there are questions raised about the later set of EVMs which are valid and need answering.

Going from that to saying the elections were rigged is a HUGE leap.

I for one think this will be "khoda pahad nikle chuiya" (dig a mountain to find a female mouse)

[Rahul Mehta]

So what is the conclusion? EVM were rigged or not?

krishnan,

I started this thread to show discuss (and show) : which mechanism paper vs EVM is less prone to mass scale rigging.

You can see that anti-RM elements have once again hijacked this thread and converted into thread with their pet themes - RM is liar, RM needs to be banned, we should do Satyagrah against RM and so forth. And a demand that thread should be killed !!

That apart,

1. Pranav has shown how a Trojan can be used to manipulate EVMs and how that Trojan would be impossible to track if the Trojan resides on the chip cache and not in the SW. (Am I right Pranav?) .

2. I have shown that someone (be CIA or anyone) with money can replace 20000 to 40000 EVM cartoons (each carton with 20 EVMs) in CEC warehouse with rigged EVM cartons in 2 weeks with 100-200 field agents and support of Chawla.

Anti-paper simply claim that large scale ballot replacement is possible, but dont say how many field agents will be needed for that. They simply claim, and give no details.

.

[Tanaji]

I do not own this site. Nor did I post even one line on that site.

So is that what you have gone down to? Quoting extracts from Pakistani web sites that claim elections are rigged? Bravo...

Again its too much to ask for actual proof. Simply stating

With this program change, an entry program is provided for any programmer to modify the PROM (Programmable Read Only memory) since time is determined outside of the balloting unit and thus beyond the control of the EVM system.

is not enough.

BTW, is that even true? You dont need a PROM for time stamps.

I see that you still have not written any post on COMPARING costs/security between EVM and paper.

True. There is a slight problem that prevents me from doing that, see? It is called ethics and being honest to your self. Unlike you, I dont pull numbers from my hat and make back of the envelope calculations and then ask others to disprove my claim. I dont have the exact knowledge as to how much paper costs, printing ink costs, labour and transportation costs etc. etc. and project it for future cases as well. So to make any useful study, one requires this type of information. I don't have it so I wont do it.

Obviously, you have access to such data. Care to tell us what you are using? For example, can you authoritatively state that for a given constituency in the last elections where paper ballots were used, what was the total amount spent on transporting ballots? Dont tell me your local bus contractor rates, I want the actual ones.

Intuition suggests that the electronic means would be cheaper, based on experiences in other fields.

[Raju]

>> Going from that to saying the elections were rigged is a HUGE leap.

Why ? we are talking 6 states and 200k questionable EVM devices. More than enough to rig an election.

[Sanku]

Intuition suggests that the electronic means would be cheaper, based on experiences in other fields.

We dont even have to rely on your intuition for the same. This has been authoritatively claimed to be true by EC over many years.

And yes, I trust EC, even with a Naveen Chawla I will trust EC 10000000000000000000000^100000000000000000000000 times more than I will trust the random rants of arbit individual -- there is something known as organizational existence and culture which is more than sum of individuals there..

[Sanku]

>> Going from that to saying the elections were rigged is a HUGE leap.

Why ? we are talking 6 states and 200k questionable EVM devices. More than enough to rig an election.

Ok I will bite, it will have to be proven
1) That all the questions that have been asked have answers which are not correct as per EC expectations
2) Those answers then must conclusively fit into and prove a clear cut rigging methodology (as you know Dileep et al have currently proven that even this is next to impossible given the randomization etc)
3) It will have to be proven that all the EVMs landed up in states and constituencies which made the difference.
4) If will have to be shown that it was done so well and so quietly that no one knew.

Also note that the crib that IIT D folks have made have been against a 2004 machine itself. So yet no one has said that 2007 machine does special things.

[Tanaji]

Was 2009 the first election where EVMs were used? I am curious why this debate is taking place right now. As per http://www.neoncarrot.co.uk/h_aboutindi ... stats.html the 2004 elections had a significant numbers of EVMs as well, so what gives?

Also note that the crib that IIT D folks have made have been against a 2004 machine itself. So yet no one has said that 2007 machine does special things.

Also note that they have NOT proven or shown how to get the compromised code onto a machine in EC control in their demonstration. Its one thing for me to write a program on my computer to delete data, another thing to get it on your computer to execute.

[Rahul Mehta]

.

Pranav,

In the article I am writing, I want to include your Trojan theory as well.

So I need some points to be explained

1. As you mentioned, the Trojan on the chip will add votes to Congress candidates

2. But how will the Trojan know that Congress candidate is no. 2 or no. 4 ? That would be input from human at the time of polling, right?

3. If so, to activate the Trojan in say N booths, I will need to send a human agent in those N booths, right?

If one person needs 10 minutes to go to a booth, input the key sequence and go to next booth, then in one day he can activate trojan in 6*8 = about 50 booths. He will need artificial skin like cover over his finger to remove the mark. Getting fake EC ID will be easy and I wont worry about it. So to rig all 1500 booths in a Constituency, I would need 75 agents. To rig 540 Constituencies over 6 different voting days, or 100 Constituencies in one voting round, I would need 7500 all over India.

So if EVM had Trojans, with 7500 field agents, I can rig ALL EVMs. And with 1500 agents, I can rig 20% of the booths and perhaps 40% of seats. Am I summing your arguments correctly?

[Raju]

Why do you need to send an agent to every booth ? Just the swing constituencies in the six states need to be influenced. that is where the anantashayanam avargal & colleagues can show their expertise.

later on these people can be rewarded with promotions to seal their lips.

[Tanaji]

No, CIA has tied up with Aliens and they can do this remotely only. No need to send 7500 agents! In fact, Chawla himself is an alien, and he is the only person required to do the rigging!

I can finally reveal it, here is an image of Chawla in his disguise and this is how he came here

Look at Chawla's image closely... I knew it! Rahul Mehta is right, thats why he gets along with Sonia Gandhi!

[Dileep]

1. Pranav has shown how a Trojan can be used to manipulate EVMs and how that Trojan would be impossible to track if the Trojan resides on the chip cache and not in the SW. (Am I right Pranav?) .

No, he didn't. Saying technically incorrect things is not proving anything.

He don't even know what a CACHE is. Do you? RM? Please google for that. Anyone with any exposure to microprocessor would be ROFL with that claim. More hilarious than the "audit the millions of transistors and interconnects".

I asked him to provide technically correct explanation about the vulnerability, and all he could come up was CACHE. I think the 12th std kids these days study what is cache, and how it works for a microcontroller.

Show the people here (who know microcontrollers) HOW. How are you going to put a trojan on the processor chip?

If at all Pranav has shown something, it would be his stupidity and arrogance.

[Tanaji]

*Deleted*

Dileep beat me to it

2. I have shown that someone (be CIA or anyone) with money can replace 20000 to 40000 EVM cartoons (each carton with 20 EVMs) in CEC warehouse with rigged EVM cartons in 2 weeks with 100-200 field agents and support of Chawla.

You have shown precisely nothing. Merely stating "replace" doesn't demonstrate the wherewithal or competence to commit the act. IF someone says he is Ron Jeremy doesnt mean that his wife will be exhausted all day.

[Dileep]

OMG, now this "inserting a trojan in the cache" is going to be quoted as a fait accompli all around, like the use of the totalizer.

[Rahul Mehta]

Admins,

Pls note that some people have resorted to posting insults, personal attacks, bringing time-space bending, bringing in aliens etc. Their goal is to kill this thread. Please do not kill this thread just because they are hell bent on ruining it.

If you kill this thread, the EVM posts will appear in other threads, as it did in past. I will not post EVM posts in other threads, but some people might. You cant make issue disappear by killing threads.

--------

http://www.india-forum.com/forums/index ... entry99684

Someone on India-Forum has posted a linkless article which says

Once the election process begins, the EC claims total transparency in all its actions. First of all the machines are taken out of storage and sent to the Districts. Thereafter, according to a Govt of India website, ‘….these machines are checked only by the engineers of the two PSUs before each election…..’

So as per this article, EVMs are shipped just a few days before poll date. So at the time of shipping EVMs from CEC warehouse to district, the field agent knew the candidate number of Congress candidate as well polling date. So a field agent can enter activation date, time as well as candidate number into the EVM using combination code. Or he can also replace cartons with real EVMs with rigged ones. Both are possible if EVMs were shipped AFTER date on which candidate numbers were given.

[Sanku]

Trojan would be impossible to track if the Trojan resides on the chip cache and not in the SW. (Am I right Pranav?) .

Dileep I took this from your post, RM actually said this? RM you claim to be CS grad from IIT and what not. I can not believe that a CS grad however old can make a statement like the one you did. Leave alone from IIT, even one who is 41 years old.

This is astounding!! Can you please tell which IIT you went to and some details to buttress your claims? Any links to any alumini site?

[Dileep]

Original post by Pranav:

For defense contractors that manufacture billions of dollars worth of top secret equipment each year, would it be that difficult to hardwire a trojan into a controller's cache? Don't think so.

Twisting by RM

1. Pranav has shown how a Trojan can be used to manipulate EVMs and how that Trojan would be impossible to track if the Trojan resides on the chip cache and not in the SW. (Am I right Pranav?) .

See how things gets twisted?

[Dileep]

A nitpick on RM's theory. EVMS are NOT kept in big cartons. They are stored in their own hard carrying case, stacked in a warehouse.

How do I know? I have seen the video of them being de-stacked from the warehouse here, in the presence of the collector. Came on news.

See, his premise that the EVMS are centrally stored in a state is blown apart, but he ignores it. This little 'inconvenience' will also be ignored.

[Rahul Mehta]

He don't even know what a CACHE is. Do you? RM? Please google for that. Anyone with any exposure to microprocessor would be ROFL with that claim. More hilarious than the "audit the millions of transistors and interconnects".

Dileep,

People do make ASIC with lots of code inside the chip. Cache or no cache, it is possible to put lot of intelligence on chip itself.

EVM has 16 keys. Say EVM uses 8057. So lets say BEL ordered some 8057 for example. Then it is possible for intel to make chips that looks like 8057 and behave like 8057 with one extra feature : when a human punches KeyN followed by keys 8, 10, 14 and 12 in that order, it will give 60% votes to Candidate no. N and 40% to the rest AFTER over 400 votes are entered. The mock use just a few votes. And so mock poll will not show this manipulation, while real poll will show it.

With naked eye, you cant differentiate between 8057 and 8057++. With equipment, perhaps you can.
But then, if person in-charge of making EVMs is co-operating , he would not bother.

And if EVMs are using complex chips, like say 80387, then a small change in the micro-code will be undetectable even with equipment or will be too too expensive.

----

Now tell us : is it impossible for Intel to make small change in 80387 so that it has an activation sequence that would change the behavior? And it is almost like usual 80387 without that activation code? Boss, anyone who is capable of building such micro-controller is capable of making such small undetectable change.

[Dileep]

And what is this thing about "modifying the PROM" to enable time stamp? Is time stored in PROM? We poor SDREs in the business thought time is kept by RTC chips, taking reference from a crystal, and the controller reading the time value through a serial protocol like I2C.

BEL has "improved" the software. That proves that it is compromised isn't it?

[Tanaji]

He don't even know what a CACHE is. Do you? RM? Please google for that. Anyone with any exposure to microprocessor would be ROFL with that claim. More hilarious than the "audit the millions of transistors and interconnects".

Dileep,

People do make ASIC with lots of code inside the chip. Cache or no cache, it is possible to put lot of intelligence on chip itself.

EVM has 16 keys. Say EVM uses 8057. So lets say BEL ordered some 8057 for example. Then it is possible for intel to make chips that looks like 8057 and behave like 8057 with one extra feature : when a human punches KeyN followed by keys 8, 10, 14 and 12 in that order, it will give 60% votes to Candidate no. N and 40% to the rest AFTER over 400 votes are entered. The mock use just a few votes. And so mock poll will not show this manipulation, while real poll will show it.

With naked eye, you cant differentiate between 8057 and 8057++. With equipment, perhaps you can.
But then, if person in-charge of making EVMs is co-operating , he would not bother.

And if EVMs are using complex chips, like say 80387, then a small change in the micro-code will be undetectable even with equipment or will be too too expensive.

----

Now tell us : is it impossible for Intel to make small change in 80387 so that it has an activation sequence that would change the behavior? And it is almost like usual 80387 without that activation code? Boss, anyone who is capable of building such micro-controller is capable of making such small undetectable change.

This is pure BS. Your usage of technical jargon does not make it true. What has 80387 microcode go to do with any of this?

[Dileep]

EVM has 16 keys. Say EVM uses 8057. So lets say BEL ordered some 8057 for example. Then it is possible for intel to make chips that looks like 8057 and behave like 8057 with one extra feature : when a human punches KeyN followed by keys 8, 10, 14 and 12 in that order, it will give 60% votes to Candidate no. N and 40% to the rest AFTER over 400 votes are entered. The mock use just a few votes. And so mock poll will not show this manipulation, while real poll will show it.

Can you post your CS degree certificate RMji?

How would the trojan knows which IO lines are used with which key, and which RAM location is used to store the vote count of each row? How does it know the entry points of the "good program" so that it can call the other functions when needed? How would you call the trojan code from the main program, without modifying the main program itself?

This idea has fallen deep below being absurd, it insults my intelligence!!

[Sanku]

BS.

This is pure BS. Your usage of technical jargon does not make it true. What has 80387 microcode go to do with any of this?

BS on unimaginable level, RM jee you wouldn't know a potato chip from a silicon one even if you ate the whole lot.

What solid BS.

I am now pretty sure that there is no Rahul Mehta who went to IITD for CS and graduated in 1990.

[pgbhat]

This is pure BS. Your usage of technical jargon does not make it true. What has 80387 microcode go to do with any of this?

IOW techno-babble onlee.

[Pranav]

I dont think anyone has demonstrated a probable series of steps that can be carried out in total secrecy.

One should talk about possibilities rather than probabilities. We already know the CEC is a compromised person. The question is whether the PSUs might be able to resist infiltration by political appointees. How many people at BEL would know if one strategically placed manager tampered with the binaries? How do the people at BEL verify the microcontrollers and other chips that they import?

[AnimeshP]

I think the discussion on this thread can be summed up in the words of John Lennon

People say I'm crazy doing what I'm doing,
Well they give me all kinds of warnings to save me from ruin,
When I say that I'm o.k. they look at me kind of strange,
Surely your not happy now you no longer play the game,

People say I'm lazy dreaming my life away,
Well they give me all kinds of advice designed to enlighten me,
When I tell that I'm doing Fine watching shadows on the wall,
Don't you miss the big time boy you're no longer on the ball?

I'm just sitting here watching the wheels go round and round,
I really love to watch them roll,
No longer riding on the merry-go-round,
I just had to let it go,

People asking questions lost in confusion,
Well I tell them there's no problem,
Only solutions,
Well they shake their heads and they look at me as if I've lost my mind,
I tell them there's no hurry...
I'm just sitting here doing time,

I'm just sitting here watching the wheels go round and round,
I really love to watch them roll,
No longer riding on the merry-go-round,
I just had to let it go.

[Sanku]

One should talk about possibilities rather than probabilities.

The problem with possiblities is that there are just so many of them. We have to work in realm of what is probable. Mind you I do not not say that MMS was foisted on the nation, I merely say that EVM rigging route is both a rather improbable and has a lower ROI than many other cleaner routes.

[krishnan]

A pinch of probability is worth a pound of perhaps

[suryag]

RMji - please dont make faluda out of your reputation. I frankly dont understand how you are coming up with this theory for one we might not even have a decoder/demux chip in the ballot/control unit. Please come up with something more constructive. IMO, this is mere trolling/flame-baiting

[Tanaji]

The only valid attack I currently see is using the totalizer.

Not because its weak, but because we don't have the details available to us yet about it. From what has appeared is that Totalizer was not used in the elections, and even when it is going to be used, it will be in those areas where there has been a history of intimidation based on booth level results. The security protocols for totalizer aren't public yet, but should be when it finds more use.

A security audit of the EVM and totalizer would be very welcome along with the protocols that EC will follow. They should post it on their website along with the other documents.

[Rahul Mehta]

This is pure BS. Your usage of technical jargon does not make it true. What has 80387 microcode go to do with any of this?

Which jargon did I use? Micro-code is not jargon anymore. Too many microcontrollers these days have micro code inside them. Micro-code in use since late 1980s or may be even before that.

-----

One should talk about possibilities rather than probabilities. We already know the CEC is a compromised person. The question is whether the PSUs might be able to resist infiltration by political appointees. How many people at BEL would know if one strategically placed manager tampered with the binaries? How do the people at BEL verify the microcontrollers and other chips that they import?

Pranav,

My question was : if chips are manufactured with flaws or features, what is one possible way by which chip maker can add a flaw or feature, and how will one who knows that flaw will use this feature or flaw at EVM in the booth?

There are several different ways of rigging. IMO, we should give description of end to end process.

----

Sanku, Dileep,

If you guys are hoping that by throwing insults on me, you can provoke me, you have wasted your time and are welcome to waste more. And you are welcome to ask any authority in IITD who roll number 86119 was in year 1986. And also feel free to call Rutgers Univ CSE dept New Brunswick. I wont waste my time anymore on discussing my degrees with you guys.

---

Dileep,

And while you throw insults, pls answer following questions

1. What equipments exist to differentiate between regular 8057 and an 8057 manufactured by Intel with a slightly different microcode? 8057 or 80387 are just examples. I dont know which chipset EVM uses. But if that chipset has microcode, my question applies on that as well.

2. Are you guys sure that BEL did use those or other equipment while manufacturing EVMs? How did YOU convince yourself that BEL did take effort to ensure that microcode inside complex chips was not changed by Intel or whichever company that manufactured it?

3. We are comparing relative fraud-less-ness of EVMs and paper ballots. So how many field agents would you need to rig N booths, N is large (in 10s of thousands or lakhs).

4. I am still waiting for cost estimates on paper ballot process from you guys. You not given even one argument or data to show that paper is MUCH more expensive than EVMs.

So I welcome you to throw as many insults as you want. But pls do answer above questions. Because it appears that you have run out of answers and so you are resorting to insults.

[Rahul Mehta]

The only valid attack I currently see is using the totalizer.

Not because its weak, but because we don't have the details available to us yet about it. From what has appeared is that Totalizer was not used in the elections, and even when it is going to be used, it will be in those areas where there has been a history of intimidation based on booth level results. The security protocols for totalizer aren't public yet, but should be when it finds more use.

A security audit of the EVM and totalizer would be very welcome along with the protocols that EC will follow. They should post it on their website along with the other documents.

Tanaji,

If totalizer is not used, how would you keep booth level counts secret? Not that I care about confidentiality, but it is one issue that many people do care. When ballot papers were used, the counting staff would open boxes of 5-10 non-adjacent booths on one table and then do counting. So not even CEC would know booth level voting for individual parties. As of today, CEC knows booth level voting and so does DC and Tahsildaar incharge of Assembly seat (ARO - Assistant Returning Officer or some such title). And CEC, RO or ARO can surely give this information to PM, CMs etc for a favor or a bribe. And if you use Totalizer, that rigging is as easy as ABC. You dont need to rig 1.4 m EVMs or replace them You only rig some 5000 to 10000 Totalizers, and bingo --- you are done.

Aside : I dont give a damn about confidentiality of voting. I am only putting this issue for those who do.