Thursday, September 16, 2010

Design an anonymous channel for a whistle-blower to give credible tip-offs

(The discussion on this social problem and the need to develop a solution below are copied from another blog I write: www.obvioustruths.blogspot.com

The technical possibilities discussed following the duplicated paragraphs are unique to this blog)

The importance of whistle-blowers

Many instances of corrupt practice and tax evasion occur with hundreds of people knowing about it. But the victims become hostages of the wrong-doer for many reasons and do not squeal. For instance, visualize a college or university charging under-the-table-fees of Rs 15 Lakhs (roughly US $ 30,000) for admitting students who cannot honestly compete with other applicants on merit. This is illegal on two grounds: it is illegal for educational institutions to demand and accept such payments; secondly, it generates wealth on which income tax has been evaded. It becomes the so-called black money. Thousands of students pay such “fees” every year. However, they are too scared to do anything about it before admission which they desire so strongly. After admission, they do not wish to expose their college or university and endanger their own education. Similarly, practically every company or organization which cheats on tax has a number of employees who know about it. But the employees are too scared to blow the whistle. They may even get killed for doing something like that. Under these circumstances, the rare whistle-blower has to be encouraged, protected and supported.
I assume that there are a few vigilance organizations (V. O.) that wish to encourage whistle blowers, and that they themselves are not corrupt! We need to create an Internet based mechanism meeting the following requirements, to enable such an organization to work efficiently.

Desired characteristics of the proposed mechanism

a) Gives the whistle blower privacy – no one should know who he is till he decides to reveal himself
b) Gives the vigilance organization some way of separating credible tip-offs from spurious reports
c) Provides for a mechanism to prove that a claimant is actually the one who had sent a particular set of anonymous messages earlier

The provision in c) is meant to enable a whistle-blower to claim a reward when his tip-off is acted upon and proves to be beneficial to society.

A few technical possibilities for implementing a solution

These items are labeled a), b) and c) to link them to the requirements given above.

a) Encourage the whistle blower to use any anonymizer (See http://en.wikipedia.org/wiki/Anonymizer ) he trusts, to protect his identity.

b) Encourage the whistle-blower to use create a file including his own postal address or something like that, add an irrelevant random text at the bottom of the file, and to compute the message-digest (See http://en.wikipedia.org/wiki/Cryptographic_hash_function ) of that file using any trusted, publicly available code. He should append the message digest of that file (alone) to every anonymous tip-off he sends, to serve as a masked-identity code. The file itself will be kept on his own PC or laptop. The random text will protect against someone trying out millions of known addresses to see if one of them has that message-digest.
The V.O. would give low credence to a tip-off from unproven sender (known only through his masked-identity code. However, if a few of his tip-offs prove to be useful, the V.O. can increase the credence it gives to tip-offs from that person).

c) The whistle-blower can surface at a time of his choosing to claim any reward that he is entitled to.

Notes:

1) Search for related information on the web. Have others published any solutions? Are you satisfied with them?

2) The mechanism described in b) is not too satisfactory. It will take a year or more for whistle-blower to earn his credibility this way. Can you improve on this?

3) Post your comments and suggestions on this blog. Together a number of us may do a better than any one of us can do individually.

Srinivasan Ramani September 15, 2010

Sunday, January 17, 2010

The Problem of Search – what is yet to be done?

The history of search engines has been a most exciting one. The development of effective search techniques will ultimately get ranked with the invention of printed books. These two, along with the development of the Internet, have made it possible for knowledge of human society to be made widely available among its members, but don’t get carried away! We have got a long way to go. The search problem shares something with the halting problem, at least in practical terms. If you spend time on searching for something, a good fraction of the time you come away with the information you wanted, but you never come away saying that the information you searched for is not there! You go away frustrated, feeling that searching is often a never-ending business. At least in terms of a human lifetime!
I had this feeling last week, searching for the information a friend abroad had asked me. He plans to visit India, SriLanka and Nepal on an Indian multiple entry visa. Someone has told him that if you enter India with that kind of visa and go to any other country, say Nepal, you need to stay out of India for at least 60 days before you re-enter India. I helped get a definitive answer to this by searching the Web for relevant information. It seems that the Government had announced this rule in Dec ’09 and later relaxed it in some way. I redoubled my effort to get a definitive answer to my friend. Can I get a reliable source’s final answer to this question? I searched the Web in different ways and then asked a renowned travel agency and sent email to an agency to which the Indian consulate outsources part of its visa application processing. The more I searched the more confused I got with contradictory information.
Why do I write this? I would like you to try and find the answer. You may be lucky in finding an answer that appears to be conclusive and reliable. Or you may not. Suppose you do find an answer that appears to be conclusive and reliable. How sure are you going to be about that answer?
Do try this experiment. You may not have a friend with this problem. But you may get good ideas on what more needs to be done to help people searching for information, particularly when they do not have the benefit of a professional in the field who knows where to look for information and what to trust.

Srinivasan Ramani