At my first job out of college in 1991, I was lucky to to work for a manufacturing company with 13 stöðum, not including its corporate HQ in New Jersey. I joined just when the company was rolling out a new ERP system. We were a small IT department of about ten people altogether, two of whom Did Not Travel. Part of the project involved replacing IBM System 36 boxes with HP hardware and HPUX. Everyone used green tubes to access the system.

The project rolls along and I’m sent down to Baltimore with a new co-worker, Jeff. Our job was to power up the Unix box, make sure the O/S was running, install the ERP system, configure the ERP, train people on the ERP and do custom work for folks on the spot. (This was a dream job, especially coming straight out of college). Before we could really get off the ground, we needed to unpack all the green tubes, put them on desks and wire them. And the best part was that we had to put the RJ11 connectors on ourselves.

For some reason that I never understood and actually never thought to ask about at the time, we had had some contracting company come along and run cable throughout the plant, but we didn’t have them put on the connectors. Svo, there was a "patch box" with dozens of of unlabeled cables in the "computer room" and these snaked around the building to various places in the building.

We worked our way through it over the course of a weekend, testing each wire, putting on a connector (making sure it was straight vs. crossed), ensuring the bit settings on the green tubes and printers were correct, labeling wires, making sure that "getty" was running correctly for each port and probably a thousand other things that I’ve suppressed since then. It all came together quite nicely.

En, there was one important cable that we couldn’t figure out. The plant in Baltimore had a relationship with a warehousing location in New Jersey. Some orders placed in Baltimore shipped out of that location. There were two wires that we had to connect to the HPUX box: a green tube and a printer. The green tube was easy, but the printer turned into a three-week nightmare.

If you don’t know it, or have suppressed it, dealing with green tubes and printers this way, there are various options that you deal with by setting various pins. 8-hluti, 7-hluti, parity (even/odd/none), probably others. If you get one of those settings wrong, the tube or printer still shows stuff, but it will be total gibberish, or it will be gibberish with a lot of recognizable stuff in between. Auðvitað, these pins are hard to see and have to be set by using a small flat-edge screw driver. And they are never standard.

We set up the first of many quick calls with the NJ guy (a grizzled computer hater who probably curses us to this day). We got the green tube working pretty quickly, but we couldn’t get the printer to work. It kept "printing garbage". We would create a new RJ11 connector, switching between crossed and straight. We would delete the port and re-created in Unix. We went through the arduous task of having him explain to us the pin configuration on the printer, never really sure if he was doing it correctly.

It’s about time to go live, everything in Baltimore is humming, but we can’t get the cursed printer up in NJ to work! We’ve exhausted all possibilities except for driving back up to NJ to work on the printer in person. To avoid all that driving, we finally ask him to fax us what he’s getting when it’s "garbage", hoping that maybe there will be some clue in that garbage that will tell us what we’re doing wrong.

When we got the fax, we immediately knew what was wrong. Sjá, our method of testing whether we had configured a printer correctly was to issue an "lp" command like this:

lp /etc/passwd

Grundvallaratriðum, we printed out the unix password file. It’s always present and out of the box, always just one page. You standard Unix password file looks something like this:

smith:*:100:100:8A-74(office):/home/smith:/usr/bin/sh
guest:*:200:0::/home/guest:/usr/bin/sh  

We had been printing out the password file over and over again for several weeks and it was printing correctly. Hins, to the end user, it was "printing garbage".

</enda>