Microkernel OS vs. monolithic OS

Andy Tanenbaum (author of the Minix OS) and Linus Torvalds (author of Linux) had this interesting debate on comp.os.minix in Jan & Feb of 1992.

Andy had a nice definition of things:

the
whole [monolithic] operating system is a single a.out file that runs in
'kernel mode.' This binary contains the process management, memory
management, file system and the rest. Examples of such systems are
UNIX, MS-DOS, VMS, MVS, OS/360, MULTICS, and many more.

In a
microkernel-based system, in which most of the OS runs as separate
processes, mostly outside the kernel. They communicate by message
passing. The kernel's job is to handle the message passing, interrupt
handling, low-level process management, and possibly the I/O. Examples
of this design are the RC4000, Amoeba, Chorus, Mach, and Windows/NT.

Minix
is micro-kernel and Linux is monolithic (at least as of 1992). It's a
very interesting discussion between two interesting guys.

Linus
brings up the issue of multithreading support in Linux vs. Minix. Andy
has a response that is funny until you realize what things were like
back in 1992:

When there is only one job active, the
normal case on a small PC, it [multithreading] buys you nothing and
adds complexity to the code.

Study: Few users reject cookies

A study by Web Side Story has found that only 7 out of 1000 Internet surfers reject or block cookies.

This
quote from Jason Catlett of the Junkbusters.com was amusing: "…
There's still the fact that when cookies are explained to [computer
users], they do not like them." I'm sure that statement is true given
the way Mr. Catlett probably explains cookies.

Browser cookies
are like sharp kitchen knives: most of the time they do a useful job,
but in the hands of the wrong person they can be use to harm others.
You may also be tempted to compare cookies to guns and the adage, "Guns
don't kill people, people kill people." But that comparison is too
extreme: improper use of guns has enormous consequence, unlike cookies.

And
the fact that this study shows only 7 out of 1000 block cookies seems
to say to me that the general public regards cookies in the same
category as knives, not guns.

DSL/Cable Modem Bandwidth

If you're not sure whether you're getting the full bandwidth you're paying for, try dslreports.com speedtest.This site http://home1.gte.net/awiner/ has good information about Windows 95/98/NT/2000 issues with high-speed internet connections. It also has good links to other resources.

My DSL bandwidth experience

When I first moved from dial-up to DSL I had some odd performance problems. Content from my ISP's own servers (ftp, pop3, mail) was high speed, but content from internet sites was consistently still at dial-up speeds. After no significant progress with tech support, I checked the comp.dcom.xdsl newsgroup and found my problem — I'd used a program called MTUSpeed to tweak some TCP/IP settings to maximize my modem dial-up. Those old settings were screwing up my DSL session. I reset to Win95 defaults, and presto, now I'm getting the full bandwidth.If you're on DSL and you ever messed with things like Max MTU size or TCP receive window, or you've used dial-up optimizing programs (like MTUSpeed), you should undo your changes to get back to the default windows settings.

BTW, if you're still on dial-up or know anyone who is, I do recommend MTUSpeed to get 25% better performance with your current modem..

Thoughts on privacy

Much of the current debate over privacy concerns focuses on paper tigers. The real problems are more about accountability and confusing identity with authorization.Businesses should be accountable for how they use my information. This lets you track down where the junk mail came from and ask them to stop. Governments should be accountable, too. This prevents "Big Brother" type scenarios. Accountability should be completely two-way: if some company knows my home address and phone number, then I should be able to easily find out the address and phone numbers of the managers & executives in the company. Basically, I'll show you mine if you show my yours. At that point I don't care if you know where I live or what my favorite color is, because I can find out the exact same information about you.

The lack of good authorization is masquerading as a privacy issue. Why are people concerned about the privacy of their Social Security Number (SSN) or their mother's maiden name? It's because of the rising occurrence of identity theft where people with bits of information like these can get credit cards in your name and run up bills in your name. They can also get things like driver's licenses which combined with credit cards let them commit crimes in your name. The problem is that prior to the information age we're living in now things like your SSN were hard for other's to discover. So companies used your SSN as a form of authorization. As far as the company is concerned, if the person on the other end of the phone knows your SSN then that's good enough for the company. The fundamental problem is things like your SSN are about identity, not authorization. Just because someone identifies them self as me (with an SSN) doesn't mean that a bank should go ahead and issue a credit card. Instead the bank should require authorization before issuing the credit card. With the free flow of information, previously obscure bits of identity like your SSN are no longer a good form of authorization. Companies should be using more secure forms of authorization like passwords and biometrics (e.g. fingerprints). At that point it doesn't matter if you know my social security number because that won't let you get a credit card in my name.

The problem with too much privacy is that it cuts both ways. When governments, companies, or powerful individuals have too much privacy it's easier for them to commit atrocious acts secure in the knowledge that their privacy means they may never be held accountable for their actions.


Many of my ideas and opinions about privacy have been influenced by David Brin. Although he's more well known as a science fiction writer, he's recently published The Transparent Society, a non-fiction book about privacy issues. You can read the first chapter online.

Thoughts on Copyright and Patents

I write software for a living which means people basically people pay me for my ideas. I'm all in favor of compensating people for their work, but I think current laws protecting big corporations rather than providing benefits to the the general public.

Copyrights and patents were intended to encourage people to create and invent when they otherwise might not. But eventually those creations are supposed to become available to the public. Lately, the view seems to be that copyrights and patents exist for the inventor, when in fact they exist for the public. The basis of copyright and patent law comes from section 8 of the US Constitution:

The Congress shall have power to promote the progress of science and useful arts, by securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries

The term "intellectual property" has arisen in the modern world. However it's a mistake to think that ideas can be treated as property in the same way that physical things are property.
I'd like to see some balance come back into copyright and patent system. Read More …