A Good Answer
Pilot: An Operating System for a Personal Computer
9/12:
What assumptions about a personal-computer OS did the authors of Pilot get right? Which did they get wrong?
Wrong:
-
1. The user of a pc would be a programmer. They assumed that a user would like to be popped directly into the debugger on a memory access error and that the user would be likely to want to configure spaces for efficiency. Programmers like to do these things, but the average user of today doesn't know anything about this.
-
2. Security would not be an issue on a personal computer. They seem to have assumed that any network a machine was on would itself be secure or protected.
-
3. A great proportion of data transfer would be done on physical media. Hence the emphasis on unique ids for each file and abstraction of the physical medium.
-
4. Pilot would be a popular operating system. Hence the concern with sharing data between other machines operating the Pilot OS.
Right:
-
1. Users would like their machines to be able to network with other machines. Hence their emphasis on networking as a first class portion of their operating system.
-
2. Users would like a keyboard, a mouse, and a screen to stare at (the higher the resolution the better, apparently).
-
3. Programmers would really appreciate the presence of high level I/O. An example is the many io classes in the Java API.
-
4. Users would like a multitasking machine. Hence the focus on synchronization of processes.
-
5. A disk failure should be at least somewhat recoverable.
-
6. Memory would have multiple levels (although they seemed to only imagine two).