A few weeks ago, I had a chance to meet a young writer from the Beijing Youth Daily, Sun Ming. A very interesting encounter. She was on a trip to the Philippines and Thailand to write an article about those countries, in preparation for the commemoration of the 60th anniversary of the end of World War II in Asia.
Meanwhile, in Ewan Silver, there’s an interesting article on “Distributed consensus and the Two Armies problem.” It’s a bit difficult to understand, but still interesting:
Q: Is it possible to reach agreement, within a finite time span, in a distributed system that has reliable processes but an unreliable method of communication?
…It is clear that no matter how many successful messages are sent between the two armies it is impossible for them both to agree that they should attack. To prove this: imagine that there is a method or protocol that guarantees agreement can be reached before the attack is due at dawn tomorrow. The fact that the protocol reaches an agreement before the dawn attack means there must be a finite number of messages in this protocol (messengers have to walk between the two camps and this takes time, even if they run VERY fast). We can therefore work out the minimum number of messages required to ensure the protocol works. We then assume that the final message in that minimum protocol fails to be delivered over the unreliable communication link Ã¢â‚¬â€œ the messenger has been captured by the city. Because this message was critical to the protocol reaching agreement (otherwise it would not be the last message in the protocol), it means the protocol has failed. It is now impossible to reach agreement.
Because this is a matter of life and death (literally) the armies will never agree to attack.
I wish I were more mathematically- or scientifically-oriented; but still, it’s intriguing to read problems like these because they indicate the joys of rational discussion.
The Marmot’s Hole N.K. has 547km of underground tunnels has an entry on North Korean tunnels, with an illustration, to boot.