Just found out today that I'm INTP (that's Myers Briggs, aka MBTI talk). Interesting in that my career councilor(sp??) would have thought I was an E rather than I, and I thought I was a J, not a P.

The answer to the first question was obvious, to me. "The only conversations you and I have ever had are about me," I said. "I'm perfectly happy and confident to talk about myself all day. But put me in a room of strangers and I won't even have the courage to introduce myself."

The second was more interesting. The way he described the difference between J and P went a little something like this: "J's primarily like to get all the info they can, make a decision, and then once that decision is made, that's that. They'll stay in a career all their life, even if they don't like it, rather than change. P's, on the other hand, are more go with the flow type of people who have no problem constantly changing, because they're always bringing in new information and re-evaluating their decision. Most P's I see have had a number of career changes."

That didn't mean much to me because apparently I'm a P but have only had one real career -- nay, passion -- in life. Namely, hacking code. I think that having primarily been in management for the past fews years I could see where the P begins to shine through. It's perfectly common to have a wonderful plan that will take 3 months, then be told to change it on the fly. I got good at that. I guess if I was more J I would have more difficulty at it. Of course, on the flip side, I also love to dream about having an idea and turning it into a startup, but never will -- because since I'm constantly re-evaluating my position, I never have the confidence in a single idea to say "Yup, this is the one I will bank it all on."

Always interesting to learn about oneself.

When my job search began in January I updated my resume and started sending it out. Then I was told by enough close friends the importance of the "seeding your resume with buzzwords" game that I tweaked and twisted and basically reworked some things to make sure that my resume was more buzzword dense.

And discovered weeks later that I was still sending out the old version. I was upset with myself.

Over the past two weeks I've been using an outplacement service (that my company is paying for). I just discovered that not only does my resume print badly, leaving half a page of whitespace on the first page, but it actually CUTS CONTENT OUT. A whole section of my resume does not print. Now I'm *really* upset. But I got an interview out of that one, so it can't have been too much of a deal breaker. But I do have a new version that prints nicely.

Just today a friend sent me his resume to pass along to a job I wasn't suited for. Which I did. Then three seconds later he said "Shoot, wrong cell phone number", corrected it and sent it to me again. So the first look at this guy that the job gets is a double resume because the first one had a mistake in it.

My outplacement guy tells me that the worst resume story he ever had was when a guy brought his in for assistance, and the admin folks reformatted it into their system and gave it back to him for proofing. He said "You got my email and phone number wrong." so they went back to the original he'd given them and it turns out he had both of them wrong on his original. His original that he had been circulating for 3 weeks. That's gotta suck.

Others?

Tax time is always scary. Last year I had a kid, my wife stopped working, and we refinanced the house (two mortgages down to one). AND, to help with the baby, I changed my exemptions to get more back in my paycheck every week. The stack of paperwork was amazing.

And we still got a tax refund that's about the same as a month's severance pay. Thank you, mortage interest!

AND, while I'm there, she says "Have you looked at sole proprietorship? Because I have a friend looking for web developers." Nice.

Tips:

• Unless you're used to doing it yourself and confident that you're not screwing yourself, pay somebody to do it like American Express. Cost me about $300, which is deductible the next year. And a pro can walk you through a variety of questions you might not have considered, in more depth than a typical software program might. I know that I was never happy with the level of detail I got from TurboTax.
• Research how your tax situation will change during unemployment. For instance I was told that if I put up a web site to market myself then I could actually deduct a portion of my ISP and cable modem costs. And that while my family computer is not deductible, the laptop that I do nothing but develop on might be (especially if I get some 10-99 income).
• If you own, refinance. If you need extra bucks, take it out now and pay off the credit cards.
• If you don't own, look into it. Hell of an investment. Big mortgages are scary but big mortgage interest deductions are cool.
• Have a kid. You know you want one. :)

So I watched E.T. this weekend for only the second time ever. The first would have been when I was 12 years old, and I cried that time. I remember it well, because I tried to convince my mother and older brother that I wasn't crying, I had merely poked myself in the eye.

Wow, this movie doesn't stand up well over 20 years. The dialogue seemed horrible. The plot jerked along, taking random turns. Sure, the idea is a classic, and it deserves the praise it gets for the quality of the story alone. But I mean, really. The mother actually leaves Drew Barrymore home alone when she goes to get Elliott. What is this kid, 5 in this movie? And I never fully understood why ET was dying in the first place, or why he suddenly gets better. Is that red thing in his chest a communicator or a health potion? Who is the "good government guy" and what's his story? How come the bad government guys never caught up to them at the end?

By my adult standards, today, I give it about a 6. As a kids' fairy tale, alot higher.

Got a magazine article proposal accepted today, I'm excited. I wrote a bunch of these a few years ago and discovered that it's easier than you might think. I just wanted to be published, wrote a letter to the editor of a new tech magazine saying "I have experience with ____, would you like an article?" and it turned into a small series. Now that I'm soon to be out of work I decided to see if it was still possible. It's harder now, but I'm happy that my proposal got accepted (no guarantee of publication, but thus far my hit rate has been excellent).

Ya can't make a living on it (unless you're a professional writer), but a few extra bucks in the pocket never hurt. Except at tax time.

Updated: I just discovered that in another sign of the unemployment times, many technical magazines are not paying cash for articles. They're claiming that just being published is compensation enough. Ug. So I guess I don't get any cool computer toys from this deal.

Just out of random curiosity, am I the only geek in the universe that watches pro wrestling? I started as a kid a good 20+ years ago and now it's gotten to the point where the whole thing is like a giant meta-joke. The fun comes in seeing how accurately you can predict everything (such as the rule that the word "tonight" is always followed by "in this very ring." Never just "in this ring", always this very ring.)

Thought of this when I was checking out Google Zeitgeist yesterday and noticed Curt Henning's name on the Top 10. Curt Henning was a pro wrestler for ages who recently died at a surprisingly young age.

"I'm lying in the middle of the ring flat on my back, and I look up to see Mick standing on top of the cage. He's supposed to rip open his shirt and there's this big heart drawn there...but there's nothing. My first thought is, Damn, he forgot the heart. My second thought is, I hope he doesn't land on my face."
-Triple H

So when you look at the "what do I want to do with my life" question, would you say that your approach to the answer is more top down, or bottom up? Do you envision the goal first, then break it down into milestones, then work on those? Or do you take stock of what you've already got and ask, "Ok, where can I go with this?" (I realize that's not the exact definition of topdown vs bottomup but I needed a catchy summary).

I've always wanted a way in Java to determine if there is "live" net connectivity on a given machine. That is, if I have a piece of code that's about to try connecting to a machine other than localhost, is there a way to determine if I should even try?

For a long time the answer was no. But JDK1.4 introduced "NetworkInterface" which has potential. I've already written a little utility that looks at a known set of interfaces (such as wlan0 and eth0 on my laptop). If either one is not null, then it assumes network is available. It seems to work correctly -- I yank out my pcmcia net card and it immediately switches to false, I put it back and it goes true again.

The problem is that doesn't work on Windows (apparently). The interface on my NT machine, "lan0", appears to always exist. And on a desktop machine this makes sense, since technically the interface does exist it's just not connected -- that's a different condition.

Is there a way in windows to "down" an interface, like "ifdown" in Linux?

I've been gainfully employed for 5 years. I'm laid off as of Q2. The last time I interviewed I remember going on like 12 interviews and getting 11 offers.

Those days are over. In the last 6 weeks I've had two interviews, and both chose "not to go forward."

What are people finding the interview is like these days? Lots of phone screens? Heavy on what we used to call the "tech out" portion? For both my interviews I got the same thing -- first, meet a boss and/or HR person who tell me about the company and the job, ask about my experience, and basically do the "feel good / personality fit" thing. In both cases I got asked back for a second. In both cases those were with techs. In the first case I met 2 Java coders who liked me, and 2 business analysts -- one liked me, one didn't, and I got shot down. In the second case I think I was misled a bit because the first interviewer kept asking me about my hands on experience and how I felt about rapid prototyping, but the second interview asked me nothing but questions on formal design methodology, then declared me "not technical enough."

Both of those leads came from my primary recruiter, who I have worked with before. I'm not just in his database, I know that he's calling places and trying to sell me. I've also answered a bunch of cold calls from headhunters who want my resume (I'm sure to just stick in the database) and never heard from any of them.

I've filled out numerous online applications, usually by emailing the HR contact a copy of my resume and a cover letter. Gotten back a handful of "We received your resume" form letters, but no callbacks yet.

monster and dice don't really have anything for me. Monster's primary role seems to be a place for headhunters to search for new resumes. Dice seems to have been stuffed by a company called KForce.

What's the interview like for people? Everybody's telling me that no matter how much it sucks, the fact is that you will be reduced to 20 buzzword checkboxes and if you've got 19 and the next guy's got 20, you lose. Yes, that sucks. I want to respond with "Who would want to work for a company that believes they can reduce an employee to 20 checkboxes?" but if it turns out that they're in the majority, well, I have bills to pay. :(

The recent slashdot story about new evidence of life on Mars yanked an old movie memory into the front of my brain. Called "Tin Man", from 1983, it's about a deaf man who has a computer that does all his speaking for him. He gets an operation to restore his hearing, and goes on to start a company that turns his computer into a true AI. I saw this movie when I was about 14, I think I probably had my first computer for a few months at that point.

There's a scene in the movie where the guy has brought his venture capitalists in to talk to the computer, to see how smart it is. One asks, "Is there life on Mars?" and the computer says "Yes." When they laugh, the inventor tells the computer to explain, and the computer proceeds to argue that there is a high probability that the spacecraft we've sent to Mars had bacteria and other single celled organisms aboard it, and likewise, there is at least some probability that some of it was able to adapt to life on Mars.

I remember it being a pretty cool AI movie for the early 80's. Sure as hell made me want to write an AI, I'll tell you that. My science fair project that year was an Eliza program that could do tricks, like if you asked it "What is the third month?" it could say "March". I baffled teachers by doing a few simple examples ("What is 3 x 9?") and then going into my big finish, typing "What do you get when you multiply the sum of every third number between 123 and 9843 by 5?" and holding my finger over the enter key, looking at the teacher and smiling as they said "No...way." Of course my parser was hardwired to look for sentences of the form ("What do you get when you X the Y of Z...") and just looking for potential parameters, such as "all" or "every Nth". LOTS of "if...then" logic :). What do you want, it was in BASIC.

I even had something taken right from Tin Man, where if you gave it a statement like "There is X", you could later ask it "Is there X?" and get back a true or false answer. This actually led to a lifelong love of natural language parsing. My science fair experiment the next year examined grammatical structure of sign language and Esperanto and argued that since a computer has no real world concepts for the words we use, it would be more appropriate to create a new language specifically for the purpose of communicating with computers. I believe these days that's called "Loglan". In college I remember speaking with one of the creators of Loglan (who had a disagreement with the other creators and went on to form Lojban which was the exact same language). I asked him if any humans could speak it, and he said no. Kinda defeated the idea.

Comparatively speaking, all TRON did was make me believe that if you started with a chess program you'd originally end up with the MCP. Although to be fair I did spent lots of time trying to write a good light cycles program. I could only ever program a defensive AI, though. I could never seem to get one that would make you crash.

Wargames just made me want to tie up the phonelines.

A recent ask slashdot asked for good examples of "real" portrayals of programmers in the movies. This morning I was reading a scifi book where the following is a glaringly obvious plot device -- software run amuck, original programmer brought in to solve the problem, has access to 90% of the source code, finds new and mysterious "compiled module" in the middle of it that he knows nothing about. Of course, he can't get at the source code for it (because the evil head of the corporation is actively hiding it), and "without the source, it's impossible to figure out what it does."

Ummmm....I guess this parallel universe doesnt have reverse engineering, emulators, that sort of thing. Oy.

Last night I had a dream that I was in an argument over the coming collapse of computer science. More specifically my argument seems to have been against patents. I was trying to argue that over the last 50 years there was enough open/free computer science out there that Don Knuth was able to write four massive volumes and still not be complete, but these days every little "innovation" that seems like common sense is being stamped with a ridiculous patent. This, in turn, causes small companies to have to work in fear that they might have inadvertantly arrived at the same (obvious) conclusions as somebody else who decided to patent it first, thus risking the company's livelihood once the lawyers get involved.

The really odd part was that I was arguing this over brunch...with Frank from Trading Spaces. What's really, really odd is that he didn't seem to care, and just kept making lesbian jokes. We were filling our plates at the buffet, and I'm pretty sure that Vern, one of the other designers, was already seated at the table.

I couldn't make this stuff up.

I saw Danny Hillis speak once and he made an interesting parallel between biology and computer science. He said, "If a human body loses an arm, it can still be saved -- it won't just die instantly. So why do we write software that just up and crashes as soon as a semi colon is out of place? We need to think about fault tolerant software a whole new way." (Paraphrased from faulty memory).

For several days this week I've been debugging my Windows machine that just stopped accessing the net for no discernible reason. I tried everything humanly possible, knowing full well the whole while that "Remove the router from the picture and reset the thing to factory defaults" was an option. But the way I saw it, that was the *last* option. My home network, to me, is a living, viable system that should be kept that way. I have no way of knowing who is accessing my web site or sending me mail, so I really want to avoid just disconnecting my servers at the first sign of trouble. It was only after days of trying everything else that I finally considered that option. Much like a human system, you want to not screw with it for as long as possible before starting to do drastic things.

Before resetting the router I contacted tech support. First thing the guy said was, "Reset the router." It would be nice to think that this was some psychic wisdom that he knew would fix my problem, but let's be real, just like the guys who work the help desk who only know how to reimage your machine, they're pretty much a one hit wonder. To them, the reset is the *first* line of defense, making them the polar opposite of debuggers like myself.

The fact that support droids only know how to reset or reinstall a machine is nothing new. I thought the biological analogy was interesting, though. Should a full system reset really be such a quick option, just because it can be? Don't your servers deserve more than that? (In fairness to the support droids who *do* know what they're talking about, I expect that a big part of their job is based on not wanting to waste time, and a system reset really does solve the problem a significant percentage of the time. Of course, my argument against that would be that a reset almost never *solves* the problem, it simply makes it go away by clearing out the variables that likely caused it.)

Random plug : In Cory Doctorow's "Down and Out" book he has a similar thought -- the doctor who only knows how to fix people by telling them to restore from a backup. When a patient comes in who says "No, find something else" the doctor gets all upset.

Always on a quest for things to keep my brain busy, I'm listening to "Prey" on disc in the mornings and hearing some fun "nano-fiction". But, it got me thinking...

Say you've got a 2d space on which you're about to drop n agents. Your goal is for them to arrange themselves into a circle, and then move as a single unit. The only way they know about their own position in space is relative to another agent that they are communicating with.

Here's what I came up with. All agents remain in a "beacon" mode until they have identified and connected with 2 and exactly 2 neighbors (via whatever communication mechanism they have). This should create a closed loop (think about it - 3 will make a triangle, 4 a square, etc...) Once you have two neighbors, calculate a delta force by attempting average the distance between your two neighbors (in other words, if you are A and your neighbors are B and C, move to approach AB=AC). Apply this delta force to any other forces or velocities currently affecting you.

The interesting "bug" in this system is that you might end up with a closed loop that does not use up all of your agents. Actually it is very likely that this will happen. You would then get several closed loops, using up all of your agents until you have So, how do you do it? An agent that is part of a closed loop can know that it is in a loop (using a simple heartbeat similar to feynman's tin soldiers). But how can they know that there is another loop that they should merge with?

Assume potential real world conditions, namely that N might have changed on you for better or worse. So you can't just count the length of the loop and see if it is What I'm envisioning (and if I lose my stupid job maybe I'll write :)) is a playground where you release a bunch of black dots that form a circle and then begin moving. At any time you can add new dots, which should get swallowed by the circle, or else remove some, in which case any holes should close up.

This is what I do with my brain when it is not otherwise full. Although I'm familiar with the concepts of swarms, emergent behavior and various "boid-like" creatures I've never sat down and tried to code them. What I'm talking about code be a known algorithm. I think the primary difference is I don't just want a mass of the things all moving together -- I want as close to a real circle as I can get.

What's the best way to react when you know you're going to get laid off? Allow me to suggest "smug apathy". That's when people who get to keep their jobs come up to you with their frustrated sighs and say "The company wants to do such and such! That'll never work!" and you smile and say "Yes, I know." It's all in the smile - try not to look too gleeful, but appropriately...well, smug.

It's quite fulfilling.