Chapter 10 — C++

I don't get the point of this particular chapter. I'll reiterate: The author does indeed like lisp a lot.
Regarding the last Hello World example, I've seen similar things with over-architecture today.

chapter 11- sys admins

Quote:sysadmins are highly paid baby sitters.

That's a funny quote. I think everyone liked it.

The filesystem at the time wasn't very solid. Most of today's filesystem have at least journaling and recovery mechanism.

Swapping is still a major problem, but at other scales.

The discussion related to centralized configuration files for all programs being bad is intriguing. In my opinion it's really not a bad thing, the other approach is to have each program containarized, à la nixOs, with their configuration along with them. If they aren't containarized it becomes a real mess.

On the last part of the chapter: Managing multiple machines is indeed hard. Today we have solutions like Puppet but I can imagine how much of a pain it could be. And, rdist doesn't seem to be solid software.

Chapter 12 - security

I'm don't buy the premise of the chapter. Many of the Unix pioneers were prominent security researchers. Ever heard of Robert Morris? (Obviously, because they mention the Morris worm)?

I also don't get what's wrong with "a few tricks in the operating system's kernel to enforce some sort of overall policy". Most OS do it that way. Still the metaphors are strong in this chapter!

Super user is a weakness, lack of encryption, and audit trails, etc.. I guess we moved a long way due to the constant arms race.

The classics (also mentioned by phillbush):

• SUID bit
  
• Having . in the PATH
  
• Using startup scripts as security holes...
  
• Fork bombs...