Even today there is much confusion. BSD predates Linux by a significant amount of time, if it hadn't been for AT&T there would be no Linux. Linus Torvalds wanted UNIX on his PC, he didn't want to pay for it, he got a copy of MINIX (part of Tannenbaum(sp?)'s operating systems fundamentals course) and ported it to his hardware. He thought it sucked. He began improving it, he and Tannenbaum got into many flame wars on the net, Linus kept releasing his work in progress, other people started contributing, eventually something usable was available, a long while later something stable was available. The Computer Systems Research Group (CSRG) at Berkeley was being disbanded, they wanted to release the Berkeley Standard Distribution (aks BSD) to the public. AT&T sued them (and BSDI and others) on the grounds that there was proprietary AT&T code in the release. (The first attempt was the BSD 4.3 release) The lawsuit took several years to play out, the result was that AT&T identified all of the code they considered "theirs". After a negotiation both sides agreed to what was the disputed code. Once settled the remaining members of the CSRG went about re-writing the code that was contested and replacing it in the kernel/system. The result was a "clean" version of BSD and it was dubbed "FreeBSD" because is was "free of encumberences" and an AT&T source license requirement. This became the basis of 4.4BSD and it spawned FreeBSD. Both BSDI and NetBSD followed similar routes (making their products "free of AT&T code.") and that is sort of where we would be today, except that the FreeBSD guys are a bit "chummy" about who can and cannot change the code in FreeBSD which increases stability but pisses some people off. A group of those people started "OpenBSD" which was, AFAICT, FreeBSD with no source control. (the bazaar model of development). Legally there is another HUGE difference between FreeBSD and Linux and it may end up killing Linux in the end. That is of coures the GNU Public License. Companies that build products out of machines running Linux are bound by the GPL to give away any code they develop that they ship as _part of Linux_. People who develop to FreeBSD are under no such restrictions. It has been hypothesized that this will lead to a host of incompatible FreeBSD variants, however my guess is that this won't be the case. It isn't the main APIs of the OS that people are changing, its things like drivers and improvements to the networking stack that people would rather keep as their "value add" as it were. On a related note, I'm currently in the process of creating a "full spec" system. I tried at one point to collect documentation on my desktop PC that was as complete as the documentation on my PDP-8/e. This was impossible due to the fact that Intel requires a signed non-disclosure agreement before they will describe the operation of their chips or CPUs. Fortunately, AMD has no such problem. I'm swapping out the Intel crap for a motherboard that has a documented chip set (MVP3) and a CPU that has a documented architecture (K6-2). The goal is to have a system where every component is completely described such that it would be possible, given time, for someone to completely "know it" at the register level. --Chuck McManis

Chuck wrote: Tony replied: Life isn't long enough to understand a Pentium-class processor at the gate level. At least, not long enough to do that and other fun things as well. Heck, I've got the prints for my KL10 and I still don't understand it to the gate level, and it's *MUCH* simpler than a Pentium. Of course, not having enough time to fully understand it is no excuse for not *wanting* the gate-level drawings. But you're not going to get those out of either Intel or AMD.

I'm still hoping to get a gate level drawin/description of a 6502. Even this modern big chips are a lot of fun to learn about - and it even has practical meaning - if you want to overclock your system (Since my 486-50 I tried to overclock all my PCs - the actual K6-366 is running at 417 MHz right now), you need these data sheeds/handbooks do develop a strategy beyond just jumpering a higher bus clock and/or change BIOS setings - from the memory control alone you can sometimes gain more power than from overclocking the CPU (Tweak Bios offers a lot of this now for everyone with a minimunm of technical knowledge without reading the manuals - but thats less fun :). Also if you look in deep, these chip(set)s are not as x86 dependant as one might assume - depending of the CPU you want to use sometimes one or two small PAL could be enough to use them - and PCI bus, memory controler, ISA bridges, or ser/par I/O keyboard and mouse are things you might need on any system. In fact, the Milan uses a PC chipset with it's 68040 CPU. I like to think (read: dream) of building an actual usable 'PC' with different (unusual) CPUs - What about a 65816 (with 32 Bit data bus) using a 486 chipset ... etc - or a design for a PC with exchangable CPUs ... just for playing around. Gruss H. -- Ich denke, also bin ich, also gut HRK

Of course I know that the 65816 has a 16 Bit data bus. It was just a nice idea of having a 65816 with a 32 Bit BIU to fit still available 486 chipsets (with PCI etc.) to get a kind of an up to date 'SuperApple' (Imagine, an Apple II type computer with an ERAZOR II graphics card :)) In fact, this idea isn't so far from reality, since WDC offers the 65816 as lib module for several design pathes. So, put it in a fast, big (F)PGA together with 32 Bit glue logic and add a last generation 486 chipset for memory and I/O controll, including ISA and PCI bus (question: is there a 486 chipset with AGP ? Just for the couriosity :). 10 MHz should be possible with no problem, more possible, but I don't know the gate structure of the 816 - 20 or more MHz are imaginable with modern (F)PGA chips. Just an idea :) Gruss H. -- Ich denke, also bin ich, also gut HRK

None. I haven't seen an NT box up for months at a time without a reboot due to something trivial like a network change that can be done with a Unix box without a reboot. Sounds reasonable... I've been using FreeBSD with 6+ months uptime on web servers (only rebooted because of equipment moves in the lab). Bill --- Bill Gates is a Persian cat and a monocle away from being a villain in a James Bond movie -- Dennis Miller bpechter@shell.monmouth.com|pechter@pechter.nws.net|pechter@pechter.ddns.org

Tony Duell wrote: I can't comment on FreeBSD, as I've never used it (I believe it's about equal to Linux and NetBSD on terms of security). However, I've been using Linux since 0.12, played with NetBSD years ago, and have been running OpenBSD for a couple months now on one of my systems. There is a program called 'saint' which is a offshoot of 'satan' I believe, that you can use to test your systems security. I ran it against my primary Linux box and then ran it against my OpenBSD server. Both running pretty much out of the box, the Linux box was running Red Hat 5.1 and I'd added a lot of stuff, but no real security changes. The OpenBSD box was either 2.3 or 2.4 and totally out of the box. 'saint' found a LOT of problems with Linux including a couple serious ones, there were two warnings on the OpenBSD box, what were they, I had telnet and ftp turned on. Linux on the other hand had all sorts of stuff turned on by default that I didn't need or want, but had just never bothered to turn off. Basically for a day-to-day workstation setup I'll run Linux, but if I want a fairly secure box, I'll run OpenBSD, if I want a really secure box, I'll use VMS. Consider this, I don't really like BSD, and am a major Linux fan! Zane | Zane H. Healy | UNIX Systems Adminstrator | | healyzh(a)aracnet.com (primary) | Linux Enthusiast | | healyzh(a)holonet.net (alternate) | Classic Computer Collector | +----------------------------------+----------------------------+ | Empire of the Petal Throne and Traveller Role Playing, | | and Zane's Computer Museum. | | http://www.dragonfire.net/~healyzh/ |

Zane wrote: You're not describing comparable things. OpenBSD is a specific BSD-based distribution. I'm sure you're probably running some specific Linux distribution, but you don't state that. Linux by itself doesn't have ANY network services turned on by default. So which Linux distribution (and what version) were you running, and what serious problems did saint find?

Oops, my mistake. I didn't read the message carefully enough. Sorry!