When you design a cryptographic system, you don't say "Believe me, it's secure", but "I did everything I could to ensure that you don't need to trust me". See for instance the Stamper FAQ (an interesting service, by the way): It would be very easy for me to try and say what a reliable and trustworthy sort of a fellow I am [...]. This would not be good enough!!!

Compare with politics, in which candidates for any sort of political office are precisely trying to convince you that they are reliable and trustworthy... Imagine how it would be if there was a politician whose main electoral promise was total transparency to ensure that you don't need to trust him...

Microblogging services such as Identi.ca (the one I use) and Twitter (the one I avoid because it is uselessly centralized) have an interesting feature: when you have an account on them and choose to subscribe to someone's feed, that person knows that you are following them.

However, I fear that people may become accustomed to this feature, and may come to believe that they will always know who is following them. Of course, this isn't true: for the many users who publish their microblogging entries to everyone, people can just follow the RSS feed without even having an account on the microblogging platform you're using.

The bottom line is that it doesn't make much sense to ask somebody to stop following your feed. The content of the feed is most likely public, and they can just silently follow you. If the problem is that you do not want to advertise them in your followers list, it's your job to manage the list properly, and remove the people you don't like without having to ask them to leave.

Sometimes, when I interact with other human beings, I realize that their way to think about the world is radically different from mine. To put it differently, I have a few weird opinions with a long chain of logical consequences which I take for granted, and I sometimes discover that people just don't think that way. I tend to be quite surprised when it happens, all the more so because superficially, other people and I tend to behave in the same way and can communicate quite easily. I'm always amazed by the fact that human beings probably all have a very different view of the world, but still manage to implement with very few glitches the interface required by society.

Hence this blog: by explaining things which seem obvious to me, and trying to find out why they seem to be so, I hope that, perhaps, if you believed them to be not obvious but plain wrong, you will think them out and realise that they could be true (because they are, since, as I perhaps forgot to mention, I'm always right). This is, of course, a reasonable goal: it is a well-known fact that people tend to change their basic mental beliefs by reading some random twenty-line ramblings.

Information should be distinguished from its support, but most people seem unaware of this distinction. Here are two examples:

Books.

When someone says that they love a book, it can indifferently refer to a sequence of letters with some interesting properties, or to a pretty object with a nice leather cover and a pleasant smell. A book with blank pages can be a very pleasant object; likewise, you can read and enjoy a literary masterpiece even if all you have is a cheap paperback edition. The advent of ebooks will probably force people to understand this difference...

Postal mail.

People indifferently use mail to move physical objects around and to transmit information. While the first activity probably won't disappear any time soon, the second one has been obsolete for quite some time now, because sending beams of light through optical fibers or pushing electrons through wires is incredibly more efficient. Yes, I am aware that the postal service has an unexpectedly large bandwidth if you use it to move hard drives around, I don't mean to criticize the rare people who might be using it in that way; the thing I find depressing is that even today, massive important services like the government still need to carry pieces of paper across the country to send me data. (I am also perfectly aware of the security and reliability shortcomings of email, thank you, but there are ways to make things work.)

Interestingly, what I'm saying doesn't apply to music and movies anymore; people seem to have understood that these things can be stored in a hard drive rather than in a pile of physical artifacts. The same should happen for paper mail and books some day...

I'm trying to find out if the following problem is already well-known in computer science...

Consider 2×N politicians and N TV channels. We would like to compute a schedule for N time slots such that:

• During each time slot, each TV channel broadcasts a live debate featuring exactly two politicians. This means that for each time slot, each politician appears on exactly one channel.
• At the end of the N slots, each politician has appeared on all channels. This means that for each politician, each channel is scheduled to exactly one time slot.
• No couple of politicians debated more than once. In other words, over all the debates which took place, each featured a different couple of politicians. [Notice that, since there are 2×N choose 2 couples of politicians and only N×N debates (one debate on each of the N channels for every of the N time slots), some couples of politicians will never debate.]

This problem is quite simple and has occurred at least once in practice (thanks Yannick!), so I'm wondering if it has already been studied. Maybe it is well-known, but under some weird name ("Martian elimination problem"? "Yannick's mother's problem"?), or maybe it is just a special case of some well-known generic problem. (I tried to find a simple graph-theoretic formulation, for instance, but failed.)

Actually, it seems that this is very similar to Kirkman's schoolgirl problem and the Social Golfer problem: to be more precise, I think that my problem is the social golfer problem for g = N, s = 2 and w = N.

In the meantime, since I wanted to find out if some solutions exist (and, if yes, what do they look like), I wrote a crappy C program to find solutions using backtracking. It turns out that this simple strategy finds solutions for N < 11 (apart from the special case with N = 2). For larger values of N, no solutions are found in a reasonable amount of time (though one would expect solutions to exist).

Here is one possible solution for N = 8. Each line is a time slot, each column is a politician, and the values in the cells are the channel on which each politician appears at each time slot. There may be a simple pattern, but I can't see it.

 0 0 1 1 2 2 3 3 4 4 5 5 6 6 7 7
 1 2 0 2 0 1 4 5 3 5 6 7 3 7 4 6
 2 1 2 0 1 0 5 4 5 7 3 6 7 3 6 4
 3 4 4 5 5 6 0 1 7 6 7 0 1 2 3 2
 4 3 5 6 7 3 1 7 6 0 0 2 2 4 5 1
 5 6 7 3 4 7 2 6 0 1 2 3 4 1 0 5
 6 7 3 4 6 5 7 0 1 2 4 1 5 0 2 3
 7 5 6 7 3 4 6 2 2 3 1 4 0 5 1 0

It would be quite funny to generate incomplete solutions to the problem (that is, grids with some cells left empty, for which we know that they can be legally completed in only one way). You would probably get something akin to sudoku...

A similar problem is discussed here (in French); search for "Quelqu'un qui assistait à cette réunion".