| Next | Previous | Contents |
What a final exam! Open sky, even if overcast and drizzly, is a relief. At least the exam is over now. It's a ridiculous way to test a person's ability in a topic, in my opinion. We had three hours for a bunch of definitions and minor calculations followed by two big programs. We were supposed, from memory and with no computer, to program Adams' method with adaptive step size control, and to produce the eigenvectors and eigenvalues of an arbitrary complex matrix, with near-equal eigenvalues being likely. He should be glad just to get the eigenvectors; there was no time to refine the subspace basis on the near-equal sets, despite the half hour extra that some class members begged. If he types in the programs and sees if they run, as he threatened to do, we're all going to get F's. This is just not the way programming is done. I don't sweat unless I press the appropriate button, but classmates' shirts are soaked, and I hope they don't catch cold in the chilly breeze off the bay. My breath reeks from stress, as does theirs, and I don't have the appetite to eat lunch. A good run may settle my stomach. Combat games at the Lion Foundation were never this stressful. Of course they also were just games.
After running out to the lagoon and back, I'm in a much better mood for lunch, though my fur is a bit damp from the winter drizzle. I have the usual bagels and rolls, but I treat myself to blueberry and strawberry jam that I bought. However horrid that exam was, it's now over, and that was my last final. I told the people at Xylogen that I would show up there at two to begin my job. I'm hoping also to cut a deal for housing, such as being allowed to put my exercise mat in a corner of the warehouse or whatever they have, because the dorms close on Friday and won't re-open until the new year.
This isn't my first time on a bus, and I've done some research on the routes. With my computer and paper notebook in my pack I ride southeast along El Camino Real and transfer at Stevens Creek Expressway, southwest. I watch the cross streets like a hawk, and manage not to miss the one I've been told. On foot in the drizzle, I march two blocks north, which brings me to the tiny lobby of a nondescript fake brick building with a discreet sign beside the door: Xylogen.
Me: Hi, I'm Simba Leones, to see Lixing Wu about a job.
Secretary: Yes, he's expecting you. You pronounce it as sh. Say shin, except it's really shing. Go in and turn right, the second door.
Me: Thanks for the tip. Bye.
The door is open and there's a clutter of several computers inside. Hardcopy is spread out on the desk and Mr. Wu is poring over it.
Me: Hi, Mr. Wu, I'm Simba Leones.
Mr. Wu: Come in and sit down. Pardon the mess. I've been looking forward to meeting you after our e-mail. With the salary all set, all that's necessary is for you to fill out an employment contract and a tax form, and get your computer account. Here are the forms. What do you want as a login ID, simba, same as at school? OK, reach over here and type your password twice.
Me: This form, I have zero dependents, right? And by lion person custom I'm married, but we're not going to make it official until we've reached majority in Utah. There's the question of getting the parents' permission, and that's a very bizarre story we don't want tangling up the state bureaucracy. I think these are all ready. There is one item; I'm going to have to arrange for housing because the dorms close end of tomorrow.
Mr. Wu: Yes, you mentioned that in your mail. Do you get along with kids and a cat?
Me: Yes, we all learned how to change a human diaper, and there were kids of all ages at the Lion Foundation. I mean human kids. Dogs don't like lion people, and the feeling is mutual, but cats tolerate us.
Mr. Wu: As they do humans. I talked with my wife, and you could stay in our rec room.
Me: Gee, thanks! That really simplifies things.
Mr. Wu: Suppose after work I'll drive you over to school to get your stuff?
Me: Wow, I was worried about how long I was going to have to ride the bus to get to your house, and if I could find it at all. I'd better get a map of the area and memorize it. I guess the next question is, where should I work?
Mr. Wu: Let's give you the server console, and if you'll find bookshelf space for those manuals you'll have room to work. You do use regular human chairs, don't you?
Me: Yes, my tail can just wrap around. But... Could I trade this chair for that one? It's a comfort to be able to stick my tail straight out the back. Please try not to step on it.
Mr. Wu: I won't. But one wheel is broken. If that doesn't bother you, you can switch chairs.
Me: No problem; the dorm chairs don't even have wheels. And I see the hub on the wall; my computer should probably have a local IP address for the duration. Standard negotiation? OK, I got the number. Do you want to fill me in on the network cleanup I'm supposed to do?
Mr. Wu: I'm supposed to be an accountant, but as you can see, I'm the de facto system administrator as well. Our net has evolved like fungus, not been designed. It's out of control, and people who need to work together aren't able to. I think you should first get an idea of what we have, then formulate a network plan, then work out and execute a procedure to get everything into compliance. We may need equipment. I'm hoping to avoid new servers or extensive rewiring, but let's set a budget of five hundred for additional hubs and minor wiring, and if you think more is needed we can talk about it. OK?
I don't need to do a psychodrama to assess the plan, but I also don't need a psychodrama to know the right answer.
Me: Sure. I have a feeling that I'd better trace, and if necessary label, the major wire runs before making any serious plans, so the present geometry can be preserved as much as possible. And I'd really like to see what shape the client machines are in. If they're that fungoid I'd better make notes on how they're configured now, to try to spot the compatibility problems you mentioned. Do you think that's a good idea?
Mr. Wu: Sounds good to me. You're not going to be unplugging stuff to trace it, are you?
Me: No; if I need to do that I'll do it in the early morning when nobody's here. I'm hoping to be able to just hand trace. Where's your ladder?
From atop the stepladder I can see the layout clearly. The building is one largish room with generic white interior paint and multiple fluorescent lamps hung from the arched ceiling. The lamp wiring is in flexible conduit, surface mounted; it's ugly but easy to install. I did it once at the Lion Foundation, but we hid ours better. I estimate that Xylogen has about the same floor area as the Lion Foundation, but closer to square. The front half is partitioned into offices and laboratories but the walls only go up 2.5 meters. The computer wires run in built-in slots on top of the walls, so tracing and rewiring will be a piece of cake. The back of the building is left open and has vats and ambiguous equipment, and stacked bags of what I suppose is a raw material. Near where I perch is a shelf at wall top level for the central hub and the external T1 modem.
Electrical wires also inhabit the wall-top slots, at least along this wall. At the Lion Foundation I once helped Mr. Rothko wire up a new machine; it takes two legs to climb a ladder safely so he was disqualified. That was where I learned about flexible conduit. Mr. Rothko explained to me why we had to drill two holes in the wall rather than just running the network connection through the electrical conduit. Somebody here hasn't had that lesson and I hope they don't learn the hard way while I have my hand on any computer equipment. If I can figure out how, I want to get the power and data wiring separated.
But my main concern is with the plan. Of course you assess the situation, make a plan, and execute it. But compliance? Mr. Wu shouldn't be thinking in those terms, particularly when explaining the assignment to the new kid who's supposed to make senior client people comply with his wonderful plan. My mental image is of a Panthera leo with his mouth open welcoming little moi to become an hors d'oeuvre. Why do I always get the tough assignments?
I've begged a sheet of stickers from the secretary, and I get started with tracing wires. Fortunately it's not like the dorm with so many machines; here, only fourteen ports on a sixteen port hub are occupied. But even so the process is tedious. Now there's a point: if the hub breaks, they're dead until they can get a new one. I would cascade five four-port hubs with one in storage as a spare. I write that in my notes to possibly suggest. Now wires one through five lead opposite the others...
At the end of wire one I find the secretary, whose name is Betty.
Betty: So, you're getting right down to work. My main complaints are that when Dr. Chang wants to call a meeting, you know, we have this software, but some peoples' schedules are all cluttered with junk, and others just can't get the program to work, so I have to go around to everyone's office anyway. You know, ask them which junk can be junked so we can have the meeting. And another thing, when my printer is out of toner I'd like to be able to print on the lab printer. And why can't I have a spare toner cartridge for it; it takes an act of God to get another when I run out.
Me: Once I've seen the lab machine I should be able to get you talking to it fairly easily. The schedule software may be a little harder. How's your internet access?
Betty: No problems. I can file our reports in Sacramento and Washington easier than I can print them on the stupid printer.
Me: That sounds like good news, something I won't have to fix. Could I check how your machine is set up? I'll need to check all the machines to find out what's blocking up the interconnection.
Filling up the schedule with junk: now that's creative use of software. I'd better tread very carefully around the people whose scheduling software doesn't work, so as to avoid giving the impression that I can fix it with one mouse click, which I probably can. Betty's machine has a simple setup which does everything she probably should be doing and no more. I'm not going to change anything except the print access. Hmm, the archive bits are set all over the place.
Me: Betty, what do you do for backups?
Betty: Well, there's this program that comes with the computer, but it's kind of hard to set up and it takes two boxes of floppies, so I really don't do it as often as I should. Do you know of a better program?
Me: In fact, I bought a book yesterday that has a chapter on backups. I haven't read it yet because I was studying for finals, but I think I'm going to make it number one on my list for tonight. I'll let you know if I come up with something.
Like backups on the server; I don't have to read the book to think of that because that's how we handled it at the Lion Foundation. I wonder what the technical issues are that might get in the way here, and how much a drive will cost if they don't already have one. Now isn't the time to chat about that with Betty. I put a sticker on her wire and mark the hub port number.
Me: OK, I think I'm done here. I've made a note about your printing problem, and the spare ink situation. I'll come back when I have a solution. See you.
Wire two leads to Dr. Chang's office. He is the president of the company. I greet him and Dr. Liang, the research director, over the office wall, but they are busy discussing some kind of deadline, so I will have to look at their machines later. I trace more wires. Wire eleven leads me to Dr. Jack Hazeltine in his small laboratory.
Jack: It's not that I don't care about the computer. I use it a lot for web searches and for keeping in contact with colleagues and finding out what other people, like competitors, are doing. But what I really need are that damned sonicator and a bigger water bath! The sonicator blew up three months ago and I've been badgering Wu for a new one ever since. I have to beg Jerry and Liang to use theirs whenever I have to sonicate cells, which is often. And I have cells in every water bath in the building. See here, I've even set up a little database and the most important field is the room number where the cells are living. We have a deadline coming up and this is no way to do research. And about the rumored network reorganization: I don't want to sound like an old codger, and, well, you know, as a lion person you're the most fascinating genetic engineering project I know of, and I really hope we can put aside our work problems and talk a little about your genome someday. But I suspect there's really nothing wrong with the network. Wu has the money locked up so tight, and I think he's moving in on the computers next.
Me: And I'm his agent. I'm going to tell you a secret. Please listen to the end before you decide to get hostile. I'm going to try to earn my pay and to do a good job on my assignment. But becoming dinner for a Panthera leo like yourself does not get the job done. Nor does it do this lion any good to get caught in the middle of a war. The way we work, we do our best to leave smiling faces behind. That means you smiling, and Wu smiling. I don't know how yet, or if I can succeed, but I'll try. I'll give you your first installment of smiles. The scheduling software: you're one of the ones that couldn't get it to work. I got Jerry to show it to me when I checked his machine. I think it's better to have a full, creative schedule like he does, but that's your choice. I need to check out your machine for connectivity problems, but it will be very difficult for me to find whatever mistake you've made that keeps the program from working. OK?
Jack: Hmmm, better than I expected. Go ahead, do your thing.
Me: The lab equipment controller board, is that hard to program?
Jack: Not if you pick one with good software. I had to fight with Wu; he wanted to buy the low-bid wonder, but I got my way on that one. I told him we absolutely have to be able to read tungsten-rhenium thermocouples, and his board can only handle iron-constantan.
Me: I don't know much about thermocouples...
Jack: Neither does he.
Me: Oh. You have quite a bit of software, and the dates on these files suggest you make good use of it. But there's plenty of disc space. You keep a clean machine. Mr. Rothko taught us to delete garbage promptly. He's kind of the equivalent of Mr. Wu for the Lion Foundation, the accountant and computer guy. I suppose this is a data directory but I see only a few archive bits. What do you do for backups?
Jack: Two sets of five floppies. One set goes home with me every night. If the main vat blows and flings oobleck all over the building, the important stuff is safe at home in my drawer.
Me: You're the only one so far to have decent backups. How would you feel about backing up across the net? I don't have exact details of what to propose, but I don't want to propose something that you're going to treat like the scheduling program.
Jack: You're serious about this.
Me: I've seen the so-called backup on the other machines. If the main vat blows, you'll be the only one able to function. That's not good for the company, and that's not good for you. Right?
Jack: Ask Wu about the last cockamamie scheme. I have a simple script that just copies everything in a few key directories onto my floppies. That's all I need. I've tried to educate people to keep their data separate from the programs but it didn't have that much effect.
Me: Actually, the machines I looked at seemed to have data and correspondence collected apart from the software, but maybe I haven't seen the worst ones yet. I'll keep in mind that you need a simple way to designate what to back up, and you don't need to back up the software itself. I'm done here, I think. Let me just put a sticker on your wire, and I'll let you get back to work.
Jack: The deadline looms. Shudder. I'll take you out to lunch, afterward. Be careful poking around the back room; let one of the techs show you what to keep your hands out of.
So, I've established a good relationship with the main opponent on the other side. I thought I had no chance; now, maybe, I have half a chance. Make that a quarter. The smiling face on Mr. Wu is going to be the hardest. I check out the lab automation in the pilot plant area. They have bags and bags of what looks like gerbil bedding; one of the kids in the dorm has an (illegal) gerbil and puts wood shavings in the cage. They fill the main vat with it, add water, and let the latest cell line try to turn it into polybutyrate plastic. One of the computers keeps track of the pressure and the temperatures various places in the vat, and controls the circulating pumps and the cooling fan. I make a note that the concentration of bacteria, and other variables, are continuously recorded in a file, but the file is not accessible on the net -- and could be.
I return to check the machines of Drs. Chang and Liang. I catch them both together in Dr. Liang's office, chattering in Chinese except the word ``deadline'' pops up in English from time to time. They assure me that internet access is excellent except that some of the remote servers are too slow. But both of them are working hard to steer the company through a minefield of technical problems, funding problems and competition, and they need up-to-date information on how the various jobs are progressing, as well as the current financial position. But when they try to set up a meeting with people to find out this information, handling the scheduling is like extracting teeth, and the software that was supposed to help with that is totally useless; half the people can't even get it to run.
Dr. Chang: Prof. Williams said you were a sharp one, no pun intended. If you succeed in only one thing, I hope it's to get that damn scheduling software to work. Or to junk it and get something we can use.
I don't need a psychodrama to know the right answer.
Me: Yes, sir. I've been looking at all the machines and already I've seen some areas of incompatibility that may be the source of the problem. I'll certainly give the scheduling issue my best try.
Both their machines are in adequate shape, except for the ubiquitous problems with backup, and except for a fairly typical collection of moldy files that should have been deleted long ago. It's about half past four. I return to the lab machine, which has its own printer, and set a software switch so it will accept print jobs from any machine in the company, same as Betty's printer will. Back in the lobby...
Me: OK, Betty, pick a one-page document to print. Now in the dialog box hit setup. See Oobleck? Select it. It's now printing on the lab printer. You'll want to change back to the local printer.
Betty: Wow, that's quick service! Could you watch the phone while I collect that? Wow!
She's back in a flash waving the printout and smiling broadly. Now that's the way to do a job. Unfortunately, that's about all I can do at a purely technical level. Betty shouldn't need to print on Oobleck.
It's time to go home. Mr. Wu takes me to the dorm to collect my stuff. I leave a note to Ken giving him Mr. Wu's address and phone number, home and work, so Ken could get in contact with me if needed. We go on the 280 freeway toward Mr. Wu's home.
Mr. Wu: You're awfully pensive. Betty seemed enthusiastic about getting her printing working right.
Me: Yes, I'm happy with that one, but I feel like I have a grenade in my pocket. Look, I have some ideas, but I really want to read in my system administration book to see what they have to say; I don't want to propose something riding home in your car that I'm going to have to un-propose tomorrow.
Mr. Wu: We're not a bunch of monsters.
Me: I know. Everyone was very nice, and cooperated in showing me their machines. I got lots of good data. But turning it into something useful, that's going to take thought. That's why I'm pensive. Sorry if I'm not good company.
Mr. Wu: Don't worry about it. Here we are; would you open the garage for me?
I manage to figure out the mechanism without opening it onto my toes.
| Next | Previous | Contents |