Nyx History: Part I, in the beginning....

THE BEGINNING | NYX IS INCORPORATED | NYX MOVES AGAIN

Nyx was originally founded by Andrew Burt at the University of Denver in Denver, Colorado. The first part of this history was written by Andrew Burt.

The early days...

About 1987, one of DU's deans, Peter Warren, convinced Zimmerman Metals in Denver to donate their aging DEC PDP 11/70 to DU (I believe they replaced it with a Microvax II). It was offered to the Math & Computer Science department. At that time, M&CS had only one departmental system, a Pyramid 90x (named Isis), used for departmental mail, research, etc., so I initially thought the PDP might make a good "toy" system for my operating systems class, so I recommended accepting the 11/70.

It arrived, all ten tons of it :-) (So I exaggerated. But the size -- about 20' long, 6' high, 3' deep -- is no joke.) It had:

• 1 Mb memory
• 3 RP06 disk drives (washing machine sized with removable packs; 166Mb each)
• 16 ports: 2 DZ-11 8-port boards
• a rack of 16 Racal-Vadic 1200 baud modems that didn't work
• a 1600 bpi tape drive
• a row of front panel switches (pink and purple!) and LED's

(Via other donations and purchases, this eventually became 3Mb memory and 5 RP06's (4 running, 1 spare).)

Initially, the 11/70 hardware was not functional. Don McCarthy, M&CS's operations manager, spent untold hours piecing it back together and making repairs. Eventually, some time in early 1988, it was functional enough to try to load an operating system. I had obtained a tape of Unix, 2.9 BSD from Berkeley, and got that going. (OS geeks will rejoice to hear it was upgraded to 2.10 in January 1989, which was much more like 4.3, whereas 2.9 was based on the older 4.2 -- i.e. Nyx was "state of the art" in software.)

I had originally intended to use the system in my Operating Systems class, but by the time the system was operational in late 1988, the OS class had already begun for the year in September, so I felt it would be more interesting to use the 11/70 as a BBS to access Internet content. I wrote a simple menu system in a few hours, one that would be obvious to users, but easy to write menus for. I did the first round in shell script (180 lines of code). A master's degree student, Dinesh Punjabi, was interested in doing a project, so I had him write a C version that strictly emulated the shell script, plus some enhancements I had been planning on. He finished it promptly, about 1000 lines of C. (I have since rewritten large parts, but I preserved his style for continuity.)

Nyx was not an entirely popular idea with the department chairman, Jim LaVita. He was afraid of (a) the liability issues raised by running a BBS (i.e., if some user loses a file, or if Nyx dies forever, will users come screaming to DU demanding that things be put right, or what about users uploading pirated software, or seeing/posting offensive things on netnews, etc.); (b) the security issue of users breaking into our other machines via Nyx; (c) the "appropriateness" issue -- should a University be engaged in running a bulletin board at all?; and (d) would it take too much of my time.

I convinced him that if we had a disclaimer saying Nyx was not a supported nor guaranteed entity, and might go away at any time, that would take care of the liability; that security was not an issue - it neither increased nor decreased it; that Nyx was educational, thus appropriate for a university; and that I would volunteer my time, so it wouldn't be done as paid DU time. He warily allowed Nyx to start; these arguments come up every so again, so far with the same results. Nyx was thus, basically, "tolerated".

When the system became operational, I brought in a Racal-Vadic 3451 1200baud modem I had, and Don found three others (the rack of modems was beyond help). Every system needs a name, and DU's convention has been to name systems after mythological figures (Persephone, Eos, Orion, Diana, Isis, etc.) I enjoy the hunt for naming a system, and I scoured my mythology books for a good name. One I'd never seen before suddenly leaped out at me -- Nyx, the Greek "spirit of the night", from which the word "night" itself is derived. Since (a) pronouncing it as "nicks" makes it rhyme with Unix; (b) BBS's typically get their heaviest traffic at night (indeed, Don says at 3am when he shuts it down for backups it sometimes has more people than during the middle of the morning); (c) it is short; and (d) nobody else in the uucp maps had taken the name already -- what could be more appropriate?

Thus Nyx was born.

I installed the Bnews software so users could read news. Nyx polled Isis, which received a full news feed from uunet, and maintained a complete copy of news (then only a few Mb/day). Transfers were at 9600 baud via a direct serial link from Nyx to Isis. (Screaming fast, in those days.)

With News, e-mail (uucp and internet, via Isis), some files to download, and user access to the shell (via a menu option and also the 'learn' program), I announced Nyx's existence to the world, quietly. I didn't want thousands of users crushing in, so I only announced it on one or two local BBSs and told various people I knew in late 1988. Users started trickling in. Usually I was the only one on, but every now and then, all four lines would be in use. People weren't exactly sure what this whole "net" thing was about or if they were even interested (this being years before the web and any kind of graphical interface -- years before even the text-based gopher and archie, the earliest versions of the web and search engine)... but people started checking it out. Being able to interact with others all over the world through netnews was the biggest draw, as was sharing free or shareware software. Since so few people had email access anywhere else, email wasn't an immediate hit -- and even when it did catch on, there was no spam!

Downloading software was always popular (in the days when malware was almost unheard of), so in June 1989 I shipped ten tapes to White Sands Missile Range in NM, home of the "simtel20" archives, and obtained a copy of their archives. (The largest of the day.) Unfortunately, the tapes were written at 6250 bpi, and the only 6250 bpi tape drive was across campus at the computing center. The tapes were written on a non-unix machine, and the local tape drive had a bear of a time just reading the tapes -- constant errors. I managed to load two of the three IBM PC tapes onto their system, then copied the data onto 10 tapes at 1600bpi, and read those in on Nyx. I also rearranged all the Simtel files into categories more to my liking. The whole process was so much work I never read in the remaining eight tapes (now quite outdated).

Lots of teething pains...

One minor problem was that the boot sequence it came with wouldn't autoboot - after a power failure or other crash, it would just sit there, waiting to be rebooted from the front panel switches (load the address of the boot ROM, tell it which device ID to boot from, and go). Eventually I wrote a boot module (PDP assembly, ugh) to fix that.

About, oh, six months after it was being used, it suffered a near fatal power supply death. Don didn't have the parts, and spending money on it wasn't an option. Don finally got it back running, but it had been down for over six months. I figured we'd have to start from scratch to build a user base, but surprisingly, users started calling pretty quickly.

We later connected five lines from the campus port selector ("the micom") to Nyx. These had the advantage that users could dial in at 2400 baud and log in; on-campus users could log in at 9600 baud. This worked - for a while. Usage got moderate (average of 1-2 people on all the time), but whenever someone started using 9600 heavily, it crashed. The DZ-11 terminal boards interrupted the CPU for every character, so it really ran "too fast" for the system to keep up. This only caused a crash, oh, once a week or so.

There were other crashes, one apparently caused by a bug in the unlink() system call code (used by the 'rm' command). It began to crash the system more and more often, as usage increased, and the amount of netnews increased. It eventually reached the point that it crashed about every eight hours from this bug (seeing maybe 3-4 users all the time). This would be about early 1990.

It had other hardware failures - memory boards went out, cpu boards went out. At one point, Nyx was down for two months. The tape drive was very unreliable at this point, and made it difficult to do backups. (The tape drive also generated a lot of heat, and couldn't be left on except when in use. The system, even without the tape, kept the room so warm that the machines crashed several times during cold weather (-20F) because the room reached over 100F! (The cooling in the machine room was comprised of two wall mount 1950's vintage air conditioners, which froze during cold weather; later one was replaced with just fan, which could be set to blow outside air into the room when cold.)

Eventually, the PDP was crashing nearly constantly as a result of bugs and hardware problems, sometimes staying down for days at a time.

Luckily for Nyx, in September 1990, M&CS moved departmental operations from the Pyramid to a network of Suns (sparcstation 1+'s with a 4/390 server). The main reason for the switch was that the Pyramid cost $25,000 per year for the maintenance contract (2% of the original cost, per month), and the cost for new equipment was about half, and provided workstations for faculty (rather than Televideo 925 terminals).

The department investigated selling the Pyramid, but found nobody interested at any decent price. I lobbied for it to become the new home for Nyx, still on the same terms as the PDP -- if it breaks, and can't be easily repaired, Nyx dies. The department chairman seriously considered just throwing out the Pyramid, but I convinced him that it was much better than the PDP and well worth keeping.

Nyx moves in with Isis...

Nyx in DU's computer room

So, by mid-October 1990, I had moved everyone off Isis onto the Suns, and moved all the user files (about 150 users) over from the PDP to the Pyramid, and the PDP was powered down. The Pyramid had several advantages over the PDP:

• Much faster: the 90x is about a 2mips machine
• More memory: 10Mb
• More disk space: 2 Fujitsu Eagle drives, 450Mb each
• More modems: the Pyramid had 4 2400 baud modems on it, which were not moved to the Sun since the Sun was reachable via DU's Cisco terminal server ("Sibyl").
• More ports: 32 ports, though only 16 support modems
• More micom lines: 11 lines from the micom were connected
• Stable operating system: Pyramid's "OSx" (version 4.1) is a "dual universe" system, providing two working environments, both a "pure" BSD 4.2 (+ some 4.3 enhancements) and system V release 2. I'd also run the system since purchased in 1983, so I knew it well, and had most software, like netnews, already installed.
• An ethernet card, allowing access to the internet "for real"

True to Murphy's Law, as soon as the system was off the maintenance contract on 12/31/90, it started breaking. One of the 16-port boards died. The other port board caused crashes if it ran any port faster than 4800 bps. One of the disk drives kicked the bucket. It started having random crashes.

Anticipating that this might happen, in October I wrote to the president of Pyramid, and requested them to donate the maintenance contract for Nyx. It wasn't until February that I heard anything, but they agreed. I offered them publicity, usage, etc.; but they agreed only on the condition that it NOT be publicized, as they didn't want every customer to come begging for free support. But, they came out and put Nyx back together.

The old 11/70 hardware was put up for sale; there were no takers, and in June 91 it was delivered to a scrap metal dealer.

Out of the Frying Pan...

Given that Nyx now had a direct link to the internet, and all users had access to the shell and all commands, instant access to 'telnet' (both in and out) spurred immense growth. The number of local logins became essentially unlimited via Sibyl, DU's terminal server. Once on, users found they could telnet to other sites, e.g., to play "muds". Further, since *anyone* could log into Nyx, under any name, the network was open to invasion from unknown users (so what else is new?).

These three factors caused the computing center to (nearly instantly) shut Nyx off from the net -- all packets to or from Nyx's ip# were discarded.

We had a "big meeting" with representatives of M&CS (me, Prof. Ottis Rechard, and Don), CaIR -- Computing and Information Resources, aka DU's computing center (Director Jim Tuccy, Asst. Director Bob Stocker, system admin Matt Brookover, network manager Byron Early), Colorado SuperNet (Director Ken Harmon, public access system project manager Dave Menges), and Westnet (Carol Ward) -- and probably some folks I've forgotten to list. The agenda was to decide under what (if any) terms M&CS ought to operate a free BBS. The issues were:

• Security - can't let "just anyone" have access to the net; should users be validated?
• Appropriate use - what should users be allowed to do with the net?
• Competition - is Nyx competing with CSN and their system that provides uucp/login/internet for $2/hr?
• Politics - should M&CS run a system that isn't only used by M&CS, or should CaIR run it?
• Liability - is DU liable (same old argument)?

Security was seen as the biggest issue. The first suggestion was that EVERYBODY who uses Nyx must prove who they are, by PHYSICALLY showing their ID at DU, before being able to use Nyx. Clearly preposterous; this would have the effect of killing Nyx, since nobody would come down to campus, before being allowed to log in, on the off chance that Nyx might be of interest to them. I lobbied for status quo -- anyone who wants a login gets one immediately, and everyone has access to the shell and network. We finally comprised when I suggested that the menus could be made sufficiently safe to prevent all but the most determined hackers from getting a shell; thus creating two classes of users - those who get no shell (hence no access to the network programs like telnet or ftp, nor access to compilers so they can't write their own telnet), and those who come to campus, show ID, sign a disclaimer form, and can then get a shell and limited access to the network.

There was still great fear that telnet access would lead to trouble, so the only approved access was to ftp. (News and mail of course, but those aren't interactive network agents.) Talk, IRC, etc. were also forbidden as "games" are not considered "appropriate use" for the internet. To ensure users with a shell and compiler access were not able to write their own telnets, the socket() system call was removed from the standard C library (not that it's impossible to write, but it should stop most trivial attempts).

Users apparently liked Nyx well enough to drive to campus and show ID: 30 in the first week (11/6/90-11/13/90), and 150 within a year. I obtained permission to also valid users who had a "valid" account on the internet already, which has led to about 300 validations. [But see below.]

The competition issue was settled easily enough - Nyx isn't guaranteed (no good for a business), and has limited uucp access (e.g., we can't supply a full feed), and CSN is fairly cheap - so Nyx isn't likely to take anyone from CSN (and, Nyx has info on CSN for those who want to learn about it).

The other arguments were not serious threats, and were mostly dropped.

Once the "secure" menu was installed, Nyx was back on the net.

Enhancement through Donations...

Once Nyx's life was no longer threatened, users began asking questions like, "Can we get a 9600 bps modem?" I, for one, intended to buy one for myself (even though DU has remained solidly in the 1980s with 2400 baud). The idea came up that we could do a public-TV-like "fund drive", where users who wanted to could send money. One user (Trygve Lode) posted that he saw a used USRobotics "dual standard" (HST and V.32) modem for $575 at a local store. We made this our goal, and the pledges started rolling in. Eventually, about 20 users contributed the money, in chunks from $10 to $200. When the pledges exceeded the magic number, the modem was bought (about mid December 91; and the pledges were received eventually to cover it).

With my success asking for donations from Pyramid, I tried it on Practical Peripherals, having recently bought one of their v.32 modems for myself. Via their public relations firm, I convinced them that Nyx was a worthy cause, and in December 91 they donated two v.32 modems ($700 each). (Out of gratitude, I also wrote a review of said modems for the Rocky Mountain News, Denver's largest paper.)

A long time news feed site, scicom, wanted to be able to transfer news faster; and in May 91, the wives (so I'm told) of the owners wanted to get into computing, and also wanted to help worthy causes - donated a Telebit T2500 modem ($1500 or so).

... and Into the Fire

Nyx had another brush with controversy, this time external to DU: A user began posting extremely anti-Semitic articles to netnews from Nyx. He began slowly, testing the waters I suppose, with what I considered incredible stupid and "national enquirer" level unbelievable articles about how lawyers and oilmen "run the world". I even rebutted one of his first postings, it was so stupid. (He never replied to what I said.) He moved on, however, to accusing lawyers, oilmen, and Jews. Some of his postings were of the foaming at the mouth type, making incredibly ludicrous claims. I won't even do them the honor of repeating them.

This was all posted internationally, mind you. I started receiving messages of the "hey, you're the sys admin, shut him off" type, with the reasons for shutting him off being excessive reposting of the same articles and denying that he posted articles that he did (he first said his account was broken into, then admitted doing it). Between e-mail to me and postings, I saw, oh, 15-20 or so different requests to shut him off; nobody cried "no, let him post". DU authorities found out, of course, and I was told to definitely shut him off (or Nyx would be shut off). I asked him to stop posting, and he complied.

Now, when the net found out he was shut off, all hell broke loose. I received over 50 messages, and saw probably double that many postings, accusing me, personally, and DU, of blatant hideous censorship. I posted the reasons, etc., and only fueled the fire.

I, personally, abhorred what he wrote, and found most of it truly preposterous (but nonetheless very offensive). It was as if someone earnestly believed that 2+2=5, and posted all sort of stupid proofs. However, I believe deeply in the right to free speech -- and that the only way to stamp out racism is to expose it to the light of reason, in the open; otherwise it festers in the dark and grows. The net, thankfully, agreed; the consensus was that he should be allowed to post (following netiquette guidelines; I would add that I surveyed numerous people I know who are Jewish, and they universally agreed). I convinced DU of this as well, and I posted a lengthy note about why he would be allowed to post.

Fortunately for all involved -- he never posted any further notes and his account expired. Nyx was still fragile. But I'm glad the outcome favored free speech, and that the "light" of the net did its job on his dark idiocy.

A similar event occurred not long after that, regarding a fellow who called himself a "Messianic Jew" and posted various articles to soc.culture.jewish, trying to "educate" the readers about Messianic Judaism. If you're not familiar with this, the synopsis is that Messianic Jews are extraordinarily disliked by non-Messianic Jews. The s.c.j readers descended on me to, you guessed it, told me to shut him off "or else". I negotiated a truce; his position was that he wanted a place to talk about his religion, and s.c.j was the most appropriate place; s.c.j readers felt he should go jump in a lake, since what he was posting was intolerable to them and didn't belong on the net anywhere. My solution was to creat alt.messianic, and he (grudgingly) agreed to move there. A true compromise, since s.c.j readers were against creating the group, and he was against leaving s.c.j.

A third, but not so near-calamity, was the breakin of some DU systems from an account on Nyx. Now, my philosophy on hackers can be read in the /.hacking file. However, as far as other DU systems go, Nyx is too fragile to be a base for hackers, so I kindly asked them to leave (and they complied).

In that I view Nyx as a wishful prototype of what personal computing could be like in the future, I'm glad Nyx weathered all these tribulations, and that what I consider the most rational solution was found to all of them (or Nyx wouldn't be here!).

More Troublemakers

In February '92, we had more trouble. First, I was called by a Detective Johnson of Scotland Yard (for real!) who was investigating some harrassing e-mail some professor in England had received from a user on Nyx. In the end, after lots of brouhaha, I ended up having to put disclaimers in all outgoing mail and news postings, saying, essentially, not to trust whatever the message says.

A second problem came when some hackers abused the shell access via an existing internet account, by using Nyx accounts to break into other sites on the network. This nearly resulted in Nyx being shut down, nearly resulted in no more shell access, and nearly resulted in no more shell access without a personal appears on campus (in that order). At this point I successfully argued that beyond a personal appearance, a notarized form or a personal check are just as secure.

Popularity

Official Nyx Trading Card from 1994

I haven't tracked Nyx usage as well as I should have, but I gather it is one of the biggest BBSs that is entirely free. To my recollection, in 10/90 there were about 150-200 "regular" users (defined as having an account that was used during the prior three months); and 1500 different logins had been created (roughly equal to the number of people who have ever tried Nyx).

Note Nyx is also used by various groups for internal communications, including the TTCC (Teaching Teachers Computer Competency) project, which connects public school teachers in the state (they dial in for e-mail, talk, news, files, etc.). The local Sun users group (SLUG) uses Nyx, as does the local ACM. (Other groups are welcome!)

By 2/91, after moving to the new system, and now being reachable by telnet, the number of users grew enormously as various postings mentioned Nyx. In 2/91, there were 900 accounts, 600 of which had been used within the previous three months; over 2000 different users had tried the system. Each day saw around 300 logins, from over 100 unique individuals, using the 12 dial-in ports (4 "micom" + 8 modems), plus some telnets.

By September 91, there were 2000 active accounts (after removing 500 users not active during the prior *two* months). There were as many as 50 new users per day, and 25-30 users logged in at any given time. Above 30, the system became so slow that the 60 second "security" time given to logging in expired before users had a chance to log in (a feature I consider to be useful - keeps the number of users at the threshold of usability!), so the number could have been higher.

By 10/91, there were approximately 2,300 user accounts (i.e., users who have logged in at least once in the last two months) of which about 1600 are "regular" (i.e., they use the system at least twice a month). Each day sees over 1000 logins, from over 400 unique individuals. At present, Nyx has 16 dial-up ports, which are usually all in use; other users telnet in (from all over the world -- Australia, Israel, Korea, etc.). Since Nyx began, over 5000 individuals have tried Nyx.

At this point, Nyx was so slow it started getting unusable, so another fund drive was undertaken, with the goal of raising money to purchase a faster machine. As of 12/1/91, $6500 was pledged from 100 users. The goal is to purchase a sparcserver II, for about $9500.

At the end of November 91, there were 2500 users who had used the system during the previous *six weeks*.

By 12/91, current users were at 3500, with 7000 total having tried Nyx.

By May 1992, we still had approximately 3500 active users, though this is likely a result of the usage restrictions (only 10 network ports, since the load was astronomical with more). The fund drive has just completed, raising almost $9,000 which is being used to purchase a Sun SparcServerII with 32Mb memory and a 400Mb SCSI system disk plus some v.32bis modems. The Pyramid will remain as a terminal and disk space server (since the ss2 will have no ports nor SMD disk controller).

A very selfish hacker, running an irc client at high priority, removing files, etc., has been a bother, even after being told to stop, but after a week of putting pressure on him he seems to have grasped the concept that some things in society are worth more than his personal pleasure.

In May 1992, I convinced the Rocky Mountain News, Denver's largest newspaper, to enter into an on-line venture whereby they will post their editorials and letters to the editor, and accept on-line replies.

September 1992: Current user count at about 7,000. Have had to impose maximum of 300k per user home directory since users abused the voluntary policy too much.

October 1992: The fact that anyone can create an account in any name, and then post and send e-mail, is causing what I call the "crank of the week" syndrome. It's been an ongoing problem, but sporadic. Now it's almost constant. I have the joy of getting calls and e-mail from other system admins, police, even the Secret Service. Most involve accounts used to send anonymous or forged messages.

November 1992: Sigh. Last straw. A vandal broke in as 'root' and "accidentally" destroyed the /etc/passwd (user list) file while I was out of town, crippling Nyx for the 8500 users in the meantime and causing me much grief. The problem is even when one finds out which account caused it, with open access, it's impossible to shut anyone out. DU is of course upset about the cranks, so DU and I have thus decided that Nyx must go to full user validation if Nyx is to survive. This means each user must prove their real identity, and no one except validated users can be allowed to send mail, post news, etc.

The problem is, the menu code, although fairly secure, is not 100% secure. There are enough ways to break out of the menus (thence to send mail, cause havoc, etc.) that the non-validated users must be given a completely secure environment until validated. Yuck. Thus is born the "preview" level of access, which is just that: A preview of Nyx; look, no touch.

Early 1993: Fund drive has enabled us to purchase a dozen new modems (AT&T Dataport v.32bis) and 2 1.8Gb scsi disks.

June 1993: After a prolonged fund drive, we reached enough to purchase a second machine, a Sun SparcServer 10, dubbed nyx10 a.k.a. Nox.

November 1993: After a year of "full validation", we have about 4,000 fully validated users. Another 6,000 or so preview accounts, which now do almost nothing (users using preview accounts as mail drops for harrasment and illicit deals caused even e-mail reading to be removed from preview mode).

February 1994: DU has reached its capacity for incoming calls, and people are getting busies (just calling DU voice numbers, that is). Naturally Nyx users are taking up 16 phone lines almost full time. Thus we have been asked to purchase & operate our own phone lines, which we will do effective August 1. In exchange, however, we get to lift the restriction on telnet, etc. We now have no external restriction on network traffic; only those restrictions we (I) choose to enforce (such as no graphics, slip, nntp, and other high volume and difficult to monitor protocols).

October 1994-February 1995: A spider, operating on Nyx under the name "spiderman", has been breaking in to many accounts, messing stuff up as root (mostly by accident, from lack of Unix knowledge). Said spider was in fact Kevin Mitnick, of "Cyberpunk" book fame, and captured by the FBI in North Carolina in February. The most humorous incident was the spiderman cartoon theme song parody he clumsily put into the login message. Other than making messes out of stupidity, he didn't do anything egregious on Nyx.

Also had our first test of the nyx free-speech policy ("Apologize, defend, or lose account"). Keith Cochran repeatedly agreed to stop harassing users on the net, and repeatedly continued to harass them. He ultimately was given the choice of stopping or losing his account, at which point he asked to have his account revoked (specifically, he said he refused to defend his actions, that his time limit had expired, and he expected me to carry out the policy).

March 1995: Ordered 8 more phone lines, and are about to purchase a new machine. Need to get entirely self-reliant for disk space and netnews service.

Second test of Nyx free speech policies. Two users repeatedly agreed to stop whining about each other, but continually baited the other party. These were James Keegan and Steve Pordon. The actual issue they griped about never seemed to be the real issue, the real issue (to me) seemed that they simply refused to leave each other alone despite saying they would quit fighting. For those who care, Pordon allegedly would attribute things to Keegan that Keegan says he didn't say; Keegan called this "forgery" and demanded that I punish Pordon. Pordon sure seems in the wrong here, right? So far so good, we have grounds for Pordon to lose his account *if* the charges are true. But Pordon defends himself, saying he's only editing Keegan's own words, it's parody, and also that he's only giving him a taste of his own medicine (implying Keegan does the same to other people). The sticky problem is that Keegan would only complain to me about it, always claiming it was illegal, that it was "forgery". Now, I back off when someone says something is not legal; so I told him that, that being his claim, he needed to take it to court. I'm not going to punish Pordon based on only *claims* of illegal activity. (Personally, I don't think Keegan had a chance of proving "forgery", since that seems to require some form of fraud, which seems lacking here.) Yet Keegan never would even comment on my request that he deal with it via legal channels. I'd ask them both to just cease, and they'd agree.

But they both would go right back at it, with Keegan whining to me every time it happened, as if quantity would make it somehow not require legal backing. So I decided that one way that would make them both stop was to say that if EITHER one did anything, I'd shut them BOTH off. So of course they both played their tune, and I sent them packing. (Keegan took a few parting shots saying I applied my ultimatum differently than I said it, but, fortunately, I'd posted it for the world to read and it was clear that I was applying it exactly as specified.) Most people agreed with my decision, indeed, wishing I'd done it long before.

Spider Attack 16-Apr-97

The first sign of trouble was on Monday, 14-Apr-97, about 1:00 PM MDT, when I noticed a command removing all files on Nox. I killed off that command (knowing it was a spider (hacker), but figuring it was perhaps accidental). I was wrong. It was malicious: Moments later, the same command was restarted, but after the system was crippled so I'd be unable to stop it. The same commands were started on all four Nyx machines. Certainly not an accident. Key files were targeted with the specific intent of crippling all the machines while the files were removed. Before I got the machines shut down, huge numbers of files were removed all four systems; none would boot, etc. This was pre-meditated and savage. "Cybercide" is a word that comes to mind.

Nyx's return to life is complicated by our backup situation: Nyx has a lot of disk space, and no tape drive of its own. We borrow one from the university, so our backups are on as-time-permits basis; many disks are not backed up at all. Thus many of the backups were months old or nonexistent.

I began the restoration process. I spent most of the rest of Monday, all day Tuesday, and Wednesday morning on this. I had most of the files restored/recreated (barring what was just plain lost forever). During this process, Nyx was broken into again (enough of the system was up so I could log in and do work) and it all started again.

Andrew Burt eventually beat off the hacker/spider and restored Nyx's files. The battle did leave holes in Nyx's OS which were to cause various crashes over the next few months.

...CONTINUED