There are two aspects to the article:
1. A security researcher arrested for possession of an EVM machine
2. The security researcher establishing the fact that EVMs in India are vulnerable to security attacks
As per 1, why is the Indian police wrong? If a stolen object is found with you, its you who are liable. Whether it was given by "unknown resources" is not important.
For 2. Is that something new? Is there any software system that is 100% secure? But don't we still use them - in banks, in flights and everywhere. The question is benefits vs risks. Now before someone argues about that here, I would say, please stop trolling and be reasonable. Indian election scenario is nowhere close to US or Europe or any developed country. If there is any real Security expert out here, he can vouch that security vulnerabilities are 90% because of humans and processes, and only 10% because of machines.
So what's the Indian voting scenario? We are talking about over 700 million voters spread across thousands of constituencies, in more than 5 phases, spread across more than 20 days, that involve more than 100 thousand administrative personals that are managed by an autonomous body called "Election Commission of India". During the election, the Election Commission has absolute powers. It can suspend bureaucrats, order an arrest, or re-voting in a particular constituency. Fortunately this is one of the few agencies in India that are widely respected (by all parties) for its non-partisan role. These EVMs just make it a bit easier for these election commission guys to administer a free and fare election. But still the entire operation is a no less a challenge...
Compare this with just 10 years back. When booth looting was a common story in some of the disconnected country side regions. You heard it right - booth looting refers to some armed men storming the election room, stamping ballet papers for their own candidates; and all it required was mere 20 minutes to alter the result. But these machines ensure that only "1 vote can be cast in 1 min". So even if there is any forced capture, in one hour these guys can cast at the max 60 votes. And one hour is enough for the security to arrive and take remedial action.
Also unlike US, the scale is huge in India. In order to really affect the elections one needs to rig a lot many machines, which is not very practical.
What these security researchers are suggesting is to dump the machines and go back to paper based ballet. What we must also consider is the cost of doing that. India spent around $1 billion for its last electronic voting based election. My question is that by spending N times the amount, and introducing paper based ballet, can someone ensure a free and fare election? If not, why rob this huge amount from poor people.