Books on operating system design and paleontology

There is the classic by Vahalia on UNIX implementation (new edition supposedly coming out ) which, is in my always humble opinion the only worthwhile modern OS textbook although the original edition of Shaw’s book is good too. It’s depressing to see that my list has not changed in 10 years – either I’m fossilized or nobody is writing good OS books (or both).

Any other suggestions?

The BSD books are nice, but more tours of BSD implementation than anything illuminating on OS design. Tannenbaum’s book on MINIX is at the same level – these books are valuable, no doubt. But most OS textbooks are vague treatises on threaded programming. Just wait until I get a free weekend and write my magnum opus: “Rob Pike was right, OS development is so over, why didn’t I go work for Google?”

Advertisements

When mathematicians get mad

From Outlaws of the Marsh, translated by Sidney Shapiro

The second [bandit] was Jiang Jing, from Tanzhou in Hunan. Originally he had studied for the imperial examinations. But when he failed to pass, he abandoned the pen and took up the sword. A clever fellow, he was a skilled mathematician. In ten thousand calculations, he was never off an iota. He was also good with a lance and staff, and a skilled tactician in battle. And so he was known as the Magic Calculator.

This inspired me to google Kenneth Krohn, the co-developer of the Krohn-Rhodes theorem which discovers an unanticipated group structure in finite state machines. I’d heard rumors that Krohn had troubles and find

The petitioner testified before the board that a Mexican national (victim) had swindled him out of a large sum of money in the early 1970’s, at a time when the petitioner was living in Mexico under the alias Henry Blackwell. The petitioner was charged with kidnapping but pleaded guilty to the lesser offenses. At his plea hearing, he admitted that he and another individual used a false identity to lure the victim to Dulles International Airport for purposes of extorting the money.(4) The petitioner admitted that his coconspirator met the victim at the airport and took him to a car “rented by [the petitioner], using [a] . . . credit card” under another false name. The victim was taken to an apartment in Maryland that had been rented and “coerced to write approximately seven letters to member of [the victim’s] family.” The letters were in the victim’s handwriting and had the petitioner’s fingerprints on them. The letters instructed various persons to dismiss a lawsuit the victim had instituted against the petitioner, and to transfer assets to the petitioner (i.e., Henry Blackwell). As part of the plea, the government agreed not to prosecute the petitioner for any other crimes related to the case and the petitioner expected to serve approximately three and one-half years in prison. The petitioner was sentenced to ten years’ imprisonment and he served approximately three years of his sentence.(5)

These felonies are very serious and were, one may fairly conclude, committed as retribution for a wrong perpetrated on the petitioner by the victim. At the time the petitioner committed the crimes, in 1973, he was thirty-five years old, and held a master’s degree and a doctorate from Harvard University, which he earned in the 1960’s. His doctorate is in mathematics. He was accomplished in his field. He coauthored a computer science theory that was published in a textbook. He had owned, and sold at a profit, his own computer business making him financially secure. He also owned property in Washington, D.C. See Matter of Prager, 422 Mass. 86, 96 (1996) (admission to bar denied despite board recommendation for admission; at time applicant committed felonies, he was mature, well educated, and possessed adequate financial resources).

Digital rights management catastrophe continues to approach

A correspondent back from a war somewhere writes:

More to the point, we had a large amount of Civvy laptops, and military systems together. There were major problems with dust, heat etc.

From what I read of the Tilt bits on the DRM in Vista, one malfunctioning peripheral could shut down the whole network.

In addition, there would be no way to reinstall the machine, as you cannot just put the disks in, you have to re-contact microsoft. This could be rather difficult in a war zone without communication. Trying to canibalise machines would have the same effect, disabling the operating system.

What I find most remarkable about DRM evolution is the lack of any discussion about liability – and since, I’m not a lawyer I mean here “moral liability”. If I can’t get to my bank records, or call the cops on skype when an armed intruder is in my home, or send an email on a military radio, or do any number of expected and perfectly legal operations because a DRM system has triggered who is to blame? Suppose XYZ Bank of NY is processing transactions on their rack of standard motherboards, and capacitor fails on an unused peripheral, triggering a chain of DRM actions that causes the hypervisor to decide that the underlying OS is not trustworthy or that the file containing some order is not permitted to be read. Who does the investor contact when his order to sell 1000000 Euros gets delayed until the rates have made it a terribly money losing trade? DRM operations are, in effect, denial-of-service attacks and the excuse is “I can deny service to prevent unauthorized use of some property”. But I’d hate to trying to justify the death of a hiker who couldn’t get to a GPS database because a DRM system was triggered.