Ketan.org

Beijingle

noreply@blogger.com (Ketan) — Tue, 19 Aug 2008 20:23:00 +0000

Beijingle, n: a short, catchy song promoting the capital of China

Camplane

noreply@blogger.com (Ketan) — Tue, 19 Aug 2008 20:23:00 +0000

Camplane, n: the aircraft that fly John McCain and Barack Obama around.

Yeah, I think it's dumb, too.

Freebase Parallax

noreply@blogger.com (Ketan) — Tue, 19 Aug 2008 20:23:00 +0000

Perhaps Freebase is something that can beat Wikipedia. Check out this video about their tool "Parallax." Parallax enables a model of browsing that automates what I have found myself doing manually. I have not yet tried it, so I can't say how it works, but they've certainly grasped a novel form of information retrieval. I'm not aware of anyone else doing something similar.

Fewer family names over time

noreply@blogger.com (Ketan) — Mon, 18 Aug 2008 15:27:00 +0000

I realized in the middle of the night that the total number of family names in the world is decreasing. I am not aware of anyone coming up with any new family names. New personal (first) names, yes, but not family names. It's also clear that some family names are dwindling. In most cultures that I am aware of, children take the family name of just one of their parents. In some cultures, children use their parents' family names to build their own compound family name, but there's a pruning process there as well, as otherwise their family names would get infinitely long. Since some names are disappearing, and no new names are appearing, the net result is fewer names over time.

None of the above seems wrong, but it sure is a weird thought to have.

Interviewer tips

noreply@blogger.com (Ketan) — Sun, 17 Aug 2008 01:35:00 +0000

I have been on the receiving end of a lot of technical interviews, and it surprises me how often companies let their employees do a bad job. I have come up with some suggestions for running such a session based on that experience. I have experienced every one of the problems that the following tips are meant to solve:

Preparation

Read the resume before the interview - you should not be studying it for the first time with the candidate sitting right in front of you. This is one of my biggest pet peeves. Waiting until the interview to learn about the candidate demonstrates a lack of seriousness and professionalism.

Be prepared. You should know in advance what questions you're going to ask, and think about various ways the conversation could go. The candidate is sacrificing a lot to be there, so don't waste the candidate's time while you struggle to think of something to talk about.

Coordinate with other interviewers so you don't make the candidate repeat himself unnecessarily.

Arranging the Interview

Let the candidate know what the schedule is in advance. The candidate should know:

how long will he need to be on site

how his time be divided

who will be meeting the candidate, and what their jobs are (not just titles)

what each part of the interview is meant to accomplish

Actually, before you tell the candidate all of that, you should make sure you yourself know all of that.

If two people will be sharing a time slot, they need to coordinate their questions so they both get the information they need without conflict.

Don't put a candidate with more than 2 interviewers at a time. For some, it may be intimidating, and it also poses a bit of a coordination challenge to get everyone's questions out of the way.

Minimize the number of on-site visits necessary. For local candidates, you can ask for two visits before making an offer. For candidates traveling more than 40 minutes, find a way to do everything in one visit. If you can't get all the necessary people squeezed in, reschedule, or redefine "necessary." Be mindful of rush hour and other timing concerns. If the candidate travels more than 100 miles, you need to compensate him for expenses. It is fine if you need the candidate to come on-site to sign offer paperwork, and the candidate can come back to if he wants to discuss an offer in person.

Consider weekend interviews. It's less convenient for your interviewers, but chances are the best candidates are very busy people. Being flexible is worth it. You can always offer your interviewing employees an extra day off to make up for it.

When setting up the interview, make sure the candidate has an immediate contact number in case they run into last-minute problems like a flat tire, getting lost, or the like.

Logistics

Someone needs to take charge of basic logistics: greeting the candidate, getting beverages and/or snacks, getting interviewers to the interview room at the right time, arranging breaks, and showing the candidate out at the end. This can be a recruiter, a hiring manager, or one of the potential co-workers, depending on who is available.

Consider recording interviews. That way, interviewers who aren't in the room for that particular session can still assess the candidate. Of course, you should get permission, and don't penalize candidates who are reluctant to be recorded.

Reserve a clean, quiet room with a whiteboard, a table, lots of paper and pens, enough chairs, and whatever supporting equipment is necessary, before the candidate shows up.

If you're going to keep the candidate in a single room for the duration, pick one with a window.

If you are going to have the candidate write code on a computer, find out in advance what their preferred editor is: emacs, eclipse, vim, netbeans, etc. If it's a common one, have it set up for them. If it's an uncommon one, ask them to bring a USB memory key with the appropriate environment already set up.

Attitude

Pay attention to the candidate, even if it's a pair interview and the other person is talking. No Blackberries allowed.

Remember that this is a personal interaction like any other, so don't be rude. It is not your opportunity to pontificate on your opinions. Don't be dismissive of what the candidate says.

Be sure to introduce yourself properly; don't just dive in.

Do not bad mouth your company, other companies, former employees, management, the recruiters who work for you, XYZ technology, or really, anything at all.

Questions

The candidate should write code. On a whiteboard is acceptable, but some things may be better done on a computer.

It is acceptable to assign "homework," some programming problem that the candidate should do on their own time. Aim for something that would take a good candidate between 45 and 90 minutes. Anything that takes more than 3 hours is too much to ask; anything that takes less than a half hour may be too easy to be useful. In my experience, these problems are usually posed with a blank slate for an answer. I suggest you consider instead giving the candidate existing code and ask them to extend it in some fashion. It is rare in the real world that you write fresh, new code, so your questions should reflect that.

Avoid brain teaser type questions. Too many of them rely on a trick, and it can often seem like you're giving a candidate a hard time just because.

Along with that, make sure it is apparent how a question relates to the qualifications and responsibilities of the job. If it's not apparent, explain the connection.

Do not ask the following:

What was a project that you worked on that posed a particular challenge?

Where do you see yourself in 5 years?

Why do you want to work here?

Rate yourself in X

Instead of the first one, ask about specific challenges posed by specific items on the resume (you read the resume, right?). Instead of the middle two, ask what the candidate wants to do. The last one is your job.

Don't ask questions that give you no useful information. For example, "what do you do when you have a technical disagreement with a co-worker?" There is only one right answer, and everyone knows what it is. The information content of any answer is zero.

The best questions are open-ended. That does not mean vague. If the candidate doesn't get your meaning, explain yourself using different words, and describe the sort of answer you are looking for.

Have more questions than your candidate has time to answer. Sometimes they've heard your questions before (perhaps even the day before), and you don't want to be caught flat-footed.

In some cases you want to spend the whole time going over just one problem. Choose a deep and subtle problem. That way you see if your candidate thinks deeply and subtly. You don't get as much information as you think from a variety of questions that you only explore shallowly. Nevertheless, as in #24, make sure you have backup questions, as if your candidate for some reason can't get over the initial hump, you may be spending a very awkward 45 minutes (or however long).

Test your questions on co-workers to make sure they're neither too hard nor too easy. It will also help your delivery; often, candidates will be unable to answer a question not because they're dumb, but because the interviewer asked it badly. In interviews, like anything else, practice is beneficial.

You should always give the candidate the opportunity to ask you questions, but don't use that as an excuse to avoid your duty to inform as well as evaluate. Anticipate common questions and answer those pre-emptively. Some of these are about the company and only need to be mentioned once. Others are per person, and each interviewer should discuss them. Imagine every candidate asks the following questions:

How big is the company?

What is the breakdown of employees?

How does the company earn its money? e.g., advertising, subscriptions, licensing fees, etc.

What has growth been like?

What exactly is your role?

How long have you been there?

What do you like about working there?

What don't you like?

Where did you work before?

After the Interview

Hiring decisions should be unanimous. If someone's opinion isn't important enough to block the hire, that person doesn't need to be interviewing. I had hiring manager willing to override his team to hire me a few years back. That is flattering, but I don't think I'd want to work for him.

Always always always contact the candidate after. Never leave anyone hanging. You also want to find out how well your company presented itself. If a candidate is leaving with a bad impression, you want to know it so you can fix it.

Let the candidate know within a week what the next step is. You don't have to commit to an offer, but you do need to let him that progress has been made. A week is the maximum; try to respond within 2 business days.

Leisure is Work

noreply@blogger.com (Ketan) — Sat, 16 Aug 2008 01:07:00 +0000

I gave my 2 weeks notice today. It was time. I'd been looking casually for over a year. It got a little more serious this summer. Due to a few organizational changes, the low-key start-up I joined was getting more and more assimilated into the bank, with the dysfunction that came with it. It was by no means intolerable, but things were getting worse, with no sign of improvement ahead.

That's not to say I was looking for an exit. I like my co-workers. We have a useful product. The benefits are pretty decent, with good insurance, 401(k) matching, flexible hours, and 4 weeks of vacation per year. The situation was more than tolerable, but unpleasant and often asinine bank actions were intruding. I noticed a real degradation in the environment from before my paternity leave until after. I didn't have to think very hard, and the problems that I had to solve had far more to do with banking procedures and bizarre external mainframe systems than any kind of computing. Plus, I was really starting to dislike our application. It may be wonderful for accountants at Fortune 500 companies, but I've gotten pretty sick of it.

You can probably skip to the end if you don't want to read a lot of boring stuff. I need to give a little bit of background to explain what happened next. I've never had much luck with finding jobs. I had 2 offers coming out of college. You can guess what my success rate was in 2001 and 2002 after getting laid off from Motive. For my job search in 2004, I contacted something like 8 companies. That resulted in 2 in-person interviews and 1 offer. My search in 2005 was much the same, which yielded the job I have for the next 2 weeks.

Last summer, I talked to about 4 companies. 2 progressed to the in-person interview stage, and I received no offers. One of them, a well-known Internet company, jerked me around for 3 months before I heard a no. I felt almost aggressively rejected in those cases. It seemed like companies liked me until they actually met me. It was not a happy experience. That was also the case with a company that I talked to towards the end of 2007.

That was what set my expectations of what kind of response to expect this time around, and how quickly. I had been accumulating a list of Austin software companies for the last year using del.icio.us. My list was up to about 90. I made a first pass and came up with about 10 companies. I narrowed that down to 3 companies that seemed really interesting.

One, let's call it Iota (names changed to protect the guilty, plus for a little suspense), was involved in a search vertical. Samba produced a tool assisting in what I consider an important but neglected software development practice. Finally, Rhubarb was a large but still rapidly-growing Internet company that had a small subsidiary that looked like it was working on interesting cloud computing things.

To those three I added three more. Kilroy I added because I happened to see their ad on Craig's List that day. Blazer I took a stab at because I happened to see on Linkedin that someone I knew had just gotten a job there. Finally, I'd been talking with an independent recruiting firm off and on for the last 3 years, and they wanted me to talk to Harmonica.

So I had a list of 6 companies to whom I sent out feelers. I figured I'd hear back from 2 of them and maybe get an offer from one. That was my experience. That's what I expected.

As the heavy-handed foreshadowing has already made clear, that is not what happened. I heard back from all 6. I passed all 6 phone screens. Kilroy I met in person first over lunch, and then never heard from again. I met Rhubarb next. They were disappointing. Their interview was disorganized, and one of the interviewers (and potential co-workers) was rude and antagonistic. I didn't need that. When the recruiter called me back, I said thanks, but no thanks. I can get abuse for free.

Things went a little better with Samba. They told me up front that the opening I had seen had already been filled, but they'd still like to talk to me in person just in case I was awesome. Yeah right. I didn't embarrass myself, though.

I had an interview with Blazer the next week. I thought it went pretty well. I certainly was impressed by the people I met. I heard back in a couple of days that they wanted me to meet a few of their executives. I had set up an interview with Iota for the next Thursday, so I set up the Blazer meeting Friday afternoon. That same week, I talked to the hiring manager from Harmonica. I figured it wouldn't hurt, since my other opportunities would collapse once they actually, you know, met me. They sounded pretty interesting, so I planned on talking to them further.

The Iota interview felt like it went well. They also seemed smart, and they had a lot of interesting things to work on. A friend of mine vouched for the head of development. I hadn't been perfect, though, so I made sure to do the homework with alacrity. The next day, I met the head of products and marketing at Blazer as well as the CFO. It wasn't much of an interview, more of a conversation to get to know each other. I was feeling pretty good about hearing from them.

When I got home, I got word from Iota: they were prepared to make an offer. I felt pretty good about that. I felt slightly less good when I saw the money. It was an improvement, but I wasn't sure it was enough improvement to justify the risks and costs of switching. They had an aggressive bonus program, but I don't think of bonuses as being real money. I told them I needed a little time, figuring I'd see what Blazer came up with.

I also noticed that an email from Samba had gotten stuck in my spam folder. They said they liked me, but since the position was filled, they wanted to know if I'd be willing to wait till the end of the year. Things were looking up, and Blazer didn't disappoint. Come Monday, they made their own offer. It was nearly identical to Iota's. That's what comes of disclosing your salary. They must use the same formulas. I had some thinking to do. They were both good companies with smart people. They each provided a service that was useful and profitable, which is a surprisingly rare combination. Plus, they were both just 3 miles away from my house.

I told both companies about the choice I was facing. I also had set up an in-person interview with Harmonica before I received either offer. The day before that interview, I talked with Iota and tried to get a little more money. I did it very badly, and consequently achieved nothing. However, the morning of my Harmonica interview, the VP of Engineering at Blazer came out of nowhere and bumped their offer by 8% to near the middle of the bonus range offered by Iota. That got my attention. Later that same morning, I had a talk with Iota. I mentioned the bump in the other offer, but mainly I just wanted to discuss how their development team worked. It was a fruitful conversation, further convincing me that they were an excellent organization.

Faced with two pretty decent offers, I was considering canceling the Harmonica interview. It just didn't seem fair to them since I was highly likely to take either Iota or Blazer. I called the independent recruiter who'd set me up with them to work through the right thing to do. She wasn't there. I figured I might as well go.

Harmonica came with two strikes against them. They were down south, a 15 mile drive compared to the 3 miles for the other guys. Plus they were already up to several hundred people. Still, they had sounded intriguing on the phone, and I foolishly thought that wasting their time was better than bailing out. So I went. And wow... Those guys were smart. They had it together. It was astonishing, considering that what they did seemed pretty unsophisticated on the surface. I felt pretty good. On my way out, I told the hiring manager my situation. I hated the idea of putting anyone under pressure, but I had a deadline.

The next day, I heard from the Harmonica recruiter. Things sounded good. At her invitation, I met the Harmonica team again at a coffee shop where they were having an off-site work day. It sounded like they were moving towards an offer, but it was Friday, and my deadline for both Blazer and Iota was Monday. My coffee shop appointment reaffirmed my reaction from the day before. They were good eggs.

Something in my conversation with Iota the previous day must have jogged something loose, because I got a call. This time, it was the CEO and co-founder. I'd met the other co-founder already, but the CEO was based in another part of the country. To my amazement, he asked me what it would take for me to commit. What? I was so surprised that all I could do was say what Blazer's raised offer was, which he immediately said he'd match, and resist his attempts to persuade me to commit right them.

That made for an anxious weekend. I had two very good offers. Iota had a clear edge over Blazer. Blazer's product was good, while Iota's was great, a real market changer. Plus, Iota had matched Blazer's base salary, plus they had an aggressive bonus program. On the other hand, Blazer had 4 weeks of vacation plus additional sick time, while Iota just had 15 days personal time off encompassing both. Then there was the big unknown with Harmonica. I didn't know what was going to happen.

I put together an elaborate spreadsheet. It had all the obvious things on it, like salary, insurance costs, etc. I also added other things, like vacation time and commute distance, trying to quantify as much as I could. I didn't know if Harmonica was going to offer, but they knew what I already had been offered, so I figured they'd match that if they offered anything. Iota had a clear edge over Blazer, but Harmonica was the big unknown. I put together an extensive list of attributes. Harmonica won almost half of the categories, while Iota won the rest. Clearly, this approach wasn't working.

Jessica realized it before I did. I was trying to find a way to make Harmonica the objective right choice. I really liked them. The people were smart and likable, the technology was interesting, and the company was exciting. But I had no idea if they were going to offer, or what. And my excessive analysis indicated that Iota would probably be better for my career in the long run. Harmonica's technology was solid, but Iota's was more on the cutting edge, and they were certainly likable enough. It was a tough decision.

See, I couldn't just go with the people I liked best. There were too many questions. I'm the sole income for a family of four, which meant the money mattered a lot more than it would have otherwise. I needed to be around as a father, which meant that an extra 45 minutes of commuting each day was actually significant. And, while I was leaning towards Harmonica, Iota was still a terrific company. Blazer was pretty good, too, but they were out of the running by now. In any other circumstances, I would have been happy to take any one of them. The problem was I could only take one. I wasn't anxious about making a mistake, because none of my choices was a bad one. What was keeping me up at night was the possibility of walking away from a fantastic opportunity.

Monday came (4 days ago) and with it, an offer from Harmonica. They made it hard for me; their offer was nearly identical to Iota's, when normalized for various factors. I went over it and over it all day. Finally, I decided to go with my gut. Iota was good. I liked them. But I liked Harmonica better. I called them up and accepted.

That wasn't the end of it. Apparently, Blazer and Iota wanted to be absolutely sure that I was absolutely sure. I heard from the VP at Blazer and the co-founder of Iota one more time. It was unreal. I was happy to explain my decision to them, but, even though it had been close, I was not going to change it. I just liked those guys too much, and I had made my commitment.

I don't tell all of this to brag (OK, maybe a little). I was massively surprised. I expected maybe one offer from the six. This was unreal. Also, I had previously thought that this was a good position to be in. That's what the few people clued in to my situation said, and no doubt you are thinking it too. No way. It was agonizing. I had trouble sleeping the whole weekend. No matter what, I was going to have to disappoint some people I'd gotten to know and like. That was not a happy prospect. If I could have cloned myself and taken all three, I would have. If I could have only talked to one of them, I would have, but my previous experience told me that would have yielded nothing.

I'm not going to reveal here who is behind the code names except for one. I don't think a gentleman tells, at least not to the whole world. I will reveal Harmonica. Go ahead, read that. Process it. I know what you're thinking. Here's the thing: the market may seem unsophisticated. It's not search. It's not heavy computing. It's not developer tools. It's even targeted at non-techies. Didn't I say that my current job wasn't a challenge? Listen: that's just your superficial impression. It's ok. It was mine as well.

Here's the thing: the technical team behind HomeAway is one of the sharpest I've ever met. Heck, we talked about tries in the interview, based on something that one of the developers had written just a short time before. These guys do software the way it should be done. Unit tests, code reviews, short iterations... It's a long list of things they do right. They do things I've read about, but I've never done before. If you're curious, look at my new job description. They operate globally, which means some serious scale. And they're seriously ambitious.

The market also is exciting for its own sake. It's a potentially huge market in the exotic field of travel. The business is strong in spite of an economic slowdown, simply because the market is so huge, and they have such potential for making it bigger. My work will be other people's leisure. It's a great opportunity, and I'm really looking forward to it.

Update: I realized I might have made it sound like I did not like the people at Iota, Blazer, and Samba. That's not at all accurate. Both Blazer and Samba had good people and appealing cultures. It was a little harder for me to get a bead on Iota's culture, but there were particular individuals there that I very much liked. In an ideal world, I would have gotten to know Iota a little bit better, but I had a limited time to make my decision, and my positive reaction to HomeAway was an immediate one.

I learned it by watching you!

noreply@blogger.com (Ketan) — Thu, 14 Aug 2008 13:41:00 +0000

Uma's sense of humor is pretty edgy.

A completely unoriginal thought re: Michael Phelps

noreply@blogger.com (Ketan) — Wed, 13 Aug 2008 15:23:00 +0000

Holy crap. How does he do that? These are the best swimmers from all over the world*, and he blows past them like they're standing still. And that 4x200m relay? That wasn't just Phelps; every swimmer on the US team was fantastic. It boggles the mind.

* Well, not really; swimming is dominated by the Western countries, Russia, and Japan, with an increasingly strong showing by China. There's no real representation from Latin America, Africa, and most of Asia. Pools are expensive.

Olympicism

noreply@blogger.com (Ketan) — Sun, 10 Aug 2008 01:41:00 +0000

Once again the Olympics rolls around with its vulgar displays of nationalism, chauvinism, tribalism, and a whole host of other -isms. People can't seem to appreciate the pursuit of excellence for its own sake; what matters is that your guy from your country wins, excellent or not. The Olympic phenomenon claims high ideals and principles, but it comes down to the same "we're better than you" attitude that characterizes so much of relations between societies.

That said, U-S-A! U-S-A! U-S-A! U-S-A!

Goldwash

noreply@blogger.com (Ketan) — Mon, 04 Aug 2008 13:17:00 +0000

Goldwash, v: to use the dazzling spectacle of the modern Olympics to obscure and distract from a record of oppression and environmental destruction.

GimmeMyData.com

noreply@blogger.com (Ketan) — Thu, 31 Jul 2008 00:28:00 +0000

That's one possible name for the Web 2.0 backup service. The .COM domain is available.

Web 2.0^2.0

noreply@blogger.com (Ketan) — Wed, 30 Jul 2008 15:06:00 +0000

One worrisome trend with the shift to web-based applications and services is that your data no longer live under your control. Companies have outages, cancel products, go out of business, etc. With local data, you can back up to physical media or use an online service. Data important to you can disappear, and you can do nothing about it. This is a problem, which means it's also an opportunity.

Any Web 2.0 startup worth using gives you access to your data. GMail supports POP and IMAP, Blogger and Facebook have APIs, and nearly everyone has some kind of RSS feed. It's too much to expect the average person to use those protocols directly. What we need is a product to do it for us, something that knows how to talk to Flickr or Tumblr or whatever and get a copy of our data somewhere safe.

This product would take one (or both) of two forms. One would be a web-based service. That's a gimme for a Web 2.0 offering. You pay them $5/month, they suck down and store your data. Naturally, they themselves would need a 2-way API. They'd have plugins for all the sites their customers use. You'd be able to view your emails or tweets or whatever on the site to make sure they're there. You'd also be able to download all that data in a single blob that you could back up yourself locally. Perhaps there would be two levels of service, one where they store your data, and the other where they merely provide a single point of access. They would also provide some way to reconstruct an account from their backups in case the original service had a catastrophic failure.

The other possible form would be as a desktop application. After all, if the goal is to protect you from failures in web-based services, a web-based service might seem beside the point. The desktop application would do exactly what the web-based service did, except the data would be stored locally. What you did from there would be your problem. You pays your moneys, you downloads your softwares. I can see a good case for either form, or even both.

If this is such a good idea, why don't I do it? Simply put, the risk is too great at this time in my life. I can't take 6 months off unpaid to work on something like this. Mortgage, kids, insurance... It's too much. However, I do know a few people (hint hint) who are ideally placed for this. I'll even try to come up with a good name for it.

The road to hell is paved with good intentions

noreply@blogger.com (Ketan) — Tue, 29 Jul 2008 12:32:00 +0000

As proverbs go, that one is truer than most. It isn't, however, all that helpful by itself. The only advice you can take from it is not to try to help anyone, which is quite a steep price. After years of meditation standing on one leg in a Bhutanese lamasery gift shop, I have discovered the key.

You cannot let yourself think good intentions permit you to do something that is otherwise wrong. The ends don't justify the means. Too many people believe otherwise. Sadly, they're also already familiar with the idea that some people believe otherwise, so they're inoculated against that truth, even with its piles of supporting evidence. I also don't think this is where most people go wrong.

The more significant and subtle point is that the value of help is determined by the recipient. Your friend who's having relationship troubles with the guy that you think isn't good enough for him? You're not helping when you advocate a break-up. I could lose a few pounds, but don't steal my Ben & Jerry's. Cleaning someone's house for them is a great favor, but if you put everything in the wrong place, you've just created more work. An overly-aggressive evangelist may think he's helping me by badgering me to Jesus, but he's really just being a pain in the ass.

That's the part that I think most people don't get. They want to feel like they're helping, even when the recipient of that "help" is worse off as a result. If you're truly sincere about helping someone, you'll make sure you're not imposing your own agenda. You'll make sure you're seeing things from their perspective. And most importantly, you'll make sure that you're actually improving the situation. Otherwise, your "help" is just an act of selfishness.

Mindset

noreply@blogger.com (Ketan) — Mon, 28 Jul 2008 21:14:00 +0000

I'm jumping this book to the front of the queue even though I have yet to post about dozens of other books I read before it. I posted a couple of times last year about the right way to praise kids. One of the researchers, Carol Dweck, whose work put out a book a couple years ago called "Mindset: The New Psychology of Success." The sub-title is hokey, but accurate.

Dweck divides the world into two types of people: those with what she calls a "fixed" mindset, and those with a "growth" mindset. Someone with a fixed mindset believes you either have it or you don't. You're either smart or you're dumb. You have natural athletic talent or you're a klutz. You and your beloved were either meant for each other, or you're wrong. Not only that, but those traits are rigid and, well, fixed. On the other hand, the growth mindset sees all those things as changeable if you have the right attitude and put in the effort.

That's what the aforementioned article about praise got to the heart of. When kids were praised for being smart, they became reluctant to stretch, because they became afraid of failure. Failure meant that they weren't as smart as they thought they were. It meant they were losers. They were defined by those failures. A more subtle consequence was a desperate need to blame their failure on something else, anything at all that could allow them to continue thinking of themselves as natural winners.

Kids who were praised for their effort, on the other hand, were primed to have a "growth" mindset. Initially, they were no more and no less successful than their fixed mindset peers. The difference became apparent over time. Whereas fixed mindset kids were reluctant to challenge themselves, the growth mindset kids actively sought out more difficult work. They may have failed just as much, but that failure did not define them. It happened, and they tried to learn from it to get better.

That, in a nutshell, is the general point that Dweck is trying to get across. It's a good thing she wrote a whole book about it, though. The fixed mindset is pernicious and insidious. It pops up all over the place, whether it's athletics, art, academics, or personal relationships (anybody know a synonym starting with 'a'?). Her chapter on counter-productive messages from parents is especially valuable. She describes in number of possible situations where parents can hamper their child's development by encouraging the perception that ability and talent are innate and unchangeable. Her dialogues are a little cheesy, but they get the point across.

Less successful is her use of well-known figures like Bill Clinton and Jack Welch. That could be my cynicism at work, though, as the only ones I had a problem with were her positive examples, often politicians and businessmen who had written best-selling autobiographies. I didn't have a problem with her mentions of Rafe Esquith and Marva Collins, two teachers who were very successful with techniques like Dwecks applied to kids others had given up on. I found the fixed mindset examples to be more effective, possibly because they were dramatic ones like Enron and Bobby Knight.

Perhaps part of the reason I found this book valuable is that it felt biographical. When I was in first grade, I had a special tutor for advanced math instruction; when I worked by myself in the library, I put up a sign saying, "Don't ask me what I'm doing because I won't tell you." That hostility is characteristic of the fixed mindset, according to Dweck. I also had the experience of coasting through high school and then hitting a wall in college. For a long time, I was reluctant to try hard things because the idea of failure was too intimidating. Now I've come to understand that the what matters isn't what you can do today, but that you do the best you can to be capable of more tomorrow.

Even if you're not me, and chances are you're not, it's still a good book to read if you ever have kids. You in your life may not have a problem, but you want to make sure you send the right messages. Even if you aren't sending the wrong messages, you have to work hard to send the right ones to compensate for our society's misguided values. Dweck suggests with some credibility that our society values natural, effortless ability, which can be pretty discouraging to anyone who doesn't measure up. She points out how the public mythology around so many so-called natural geniuses like Michael Jordan or Thomas Edison fails to mention their tremendous dedication. Teachers and managers would also benefit; it's disturbing how many teachers give up on many students practically the moment they meet them. It's not just about identifying these negative attitudes. Dweck discusses how to change these attitudes, both in yourself and in others. It's a difficult task requiring constant attention, but it can be done.

This is more than self-help babble. It seems like every day there's a new discovery attesting to the plasticity of the human mind. You can be better at everything you do, but only if you're willing to try. In a couple hundred pages, Dweck ably describes her findings, supports them with references to academic studies, narrates illustrative anecdotes, and provides a prescription for the reader. It's hard to ask much more.

0345472322
978-0345472328

For Those about to Rock, We Salute Wikipedia

noreply@blogger.com (Ketan) — Mon, 28 Jul 2008 16:37:00 +0000

Heavy Metal Umlaut is my favorite Wikipedia article. None of the Wannapedias I mentioned has anything about it. Not on Conservapedia, Mahalo, Knol, Citizendium, or Squidoo. Final score? Wikipedia: awesome. Everyone else: l4me.

Cheap sound-proofing

noreply@blogger.com (Ketan) — Sun, 27 Jul 2008 16:35:00 +0000

With a noisy 3-year old in the house, I spend a lot of time thinking about sound proofing. A lot of time. It's pretty expensive to get it done right, but I think I have an idea for doing it on the cheap: bubble-wrap. I gather that effective sound insulation is like insulating for temperature. You want to avoid solid surfaces touching each other and use lots of layers. Bubble wrap seems perfect. All those cells of air and layers of plastic. I figure a few sheets ought to do a pretty good job. I haven't gotten desperate enough to try it yet, or maybe I just haven't hit on the right way to do it.

Prius cost increase

noreply@blogger.com (Ketan) — Sat, 26 Jul 2008 20:09:00 +0000

Toyota's $500 increase in the price of the Prius is an important reminder that the gasoline use is only part of the energy consumed. Quoth the spokesman: "Almost everything is made out of petroleum. Rubber, plastic, transportation (costs), glass, things like that." A hybrid will save some energy, but it's important to consider where you're starting. If you have, say, a 5-year old conventional Honda Civic, a Prius will likely be a net loss, energy-wise.

Of course, that assumes that you accept Toyota's stated reason. Seems more likely that the price increase is because sales are through the roof. Claiming it's due to transportation is no doubt partly true, but also a more acceptable justification for a price increase in economic hard times.

In denial about Wikipedia

noreply@blogger.com (Ketan) — Sat, 26 Jul 2008 12:06:00 +0000

5 years ago, you would have been completely justified in being skeptical about Wikipedia. About.com was an example of what was considered better. It was controlled. The authors and editors were qualified and authoritative. Wikipedia seemed like a ridiculous exercise in utopian naïveté.

7 years into the experiment, it's clear that Wikipedia is a success. Wiki works. What was healthy skepticism increasingly looks like denial. Mostly it's just verbal sniping, but there are a number of projects that attempt to "fix" Wikipedia by restricting authorship to appropriately qualified "authorities."

Citizendium requires real names and requires all articles be approved by their group of "experts." They're not different enough nor do they have sufficient critical mass to catch up. They don't appreciate what it is that has made Wikipedia such an unstoppable force.

Conservapedia aims to fix Wikipedia's "liberal bias." What they will do about reality's known liberal bias remains unclear. Their misguided aims doom them from the start; I imagine it will be no more successful than Air America Radio.

Mahalo tries to avoid authorship altogether. They provide no information directly. It's like they took the corresponding Wikipedia article and stripped out everything but the "References" section. Their compelling advantage is thus that they provide less information and require you to do more work. Good luck with that.

Finally, Google has entered the arena with Knol, which encourages anyone to post an article on anything, and let the search algorithm sort out who's best. I thought that approach was useless when it was called Squidoo; apparently Google likes the idea, as well as the idea of giving it a dumb name. Assuming people use it, you'll end up with dozens of different pages on a single subject, all incomplete in different ways. You also cannot make a small edit to an existing article; you have to build the whole thing yourself. Just like one feature does not make a business, a single correction does not make a useful Knol.

The one fly in the ointment for Wikipedia is Google's control of the dominant search engine. Thus far, they've been true to their "don't be evil" motto in their index; if their Wikipedia jealousy causes them to pervert their search results, people will just stop using Google and go directly to Wikipedia. I already do that for many of searches (Firefox's keyword searches are indispensable). And that doesn't even get into the shrieks of delight that will echo from Redmond (Microsoft) and Sunnyvale (Yahoo) as Google's competitors contemplate what the DoJ will do in response.

Every one of these alternatives throws the baby out with the bathwater. Wikipedia's genius is how it manages to be both centralized and decentralized at the same time. What's centralized is the collection of information, something Squidoo, Mahalo, and Knol cannot match. Besides the obvious advantage of having a one-stop shop, they ignore the importance of the snowball effect; to produce a better lens or Knol, you have to start from scratch. I would be surprised if any of my regular readers has never even once made an edit to Wikipedia.

The decentralized part of Wikipedia is of course the army of authors and editors, namely, the entire population of Internet users. It turns out that really matters. Conservapedia rejects anyone who doesn't share their set of biases (i.e., most of everyone) while Citizendium puts barriers in front of anyone who might want to contribute. If there's anything we've learned in 15 years of the Web, it's that everything you put between a user and an action, no matter how seemingly trivial, chips away at the number of people who will actually bother.

These sites may improve on Wikipedia in certain narrow ways, but those improvements come at such a cost that none of them will be able to defeat Wikipedia. Wikipedia certainly could stand a few improvements. Vandalism is generally contained, but still occurs frequently. The cabal of moderators occasionally gets unhinged. Wikipedia's markup has gotten increasingly complicated as the project has adopted more sophisticated conventions for formatting and organization, restricting what a casual editor can do. Those are all real problems that Wikipedia has yet to solve. Some day, something better will come along, but these guys ain't it.

Spam suspect identification

noreply@blogger.com (Ketan) — Fri, 25 Jul 2008 13:16:00 +0000

One of the advantages of having my own domain is that I can give every service I sign up with a different, unique email address. As a result, I know that Macy's gave my address out so I could get spam from the Atlanta Journal-Constitution. Thanks!

The Peacock's Tailpipes

noreply@blogger.com (Ketan) — Wed, 23 Jul 2008 16:06:00 +0000

A luxury car is a way to advertise earning power. Only people with a lot of money can afford to waste so much of it on something so useless. It's like the peacock's tail; only the supremely fit can afford to expend so many of their resources in such a display.

Much like the peacock, luxury cars' tails display wastefulness. Cheap cars get good mileage. Luxury cars get bad mileage. More fuel consumption means more exhaust. Ordinary cars have a single, smaller tailpipe. Some pickup trucks have two, or one larger one. Luxury cars usually have 2 and often have 4. They advertise that their owners don't need to worry about the cost of gasoline; they can burn as much as they want.

We as a society conserve only because we have to. Conservation is otherwise contrary to our values. Luxury cars demonstrate what we truly aspire to, and that is to waste. Waste is the ultimate luxury. One day, perhaps there will be high-mileage luxury car. When that happens, you'll know that we've found something even better to waste. I hope it's something we can afford.

Addendum: I am aware of high mileage exotic cars like the Tesla. Those exist in such small numbers that they don't matter.

The AMA on home births

noreply@blogger.com (Ketan) — Mon, 21 Jul 2008 15:30:00 +0000

The American Medical Association has issued a statement against home births (original Word doc converted to HTML by Google). To me this provides further evidence that, at the very least, the AMA does not have the best interests of the general public at heart.

One concern is their apparent disregard for quality of life issues. Their focus is narrow, defining success as survival of mother and child:

"An apparently uncomplicated pregnancy or delivery can quickly become very complicated in the setting of maternal hemorrhage, shoulder dystocia, eclampsia or other obstetric emergencies, necessitating the need for rigorous standards, appropriate oversight of obstetric providers, and the availability of emergency care, for the health of both the mother and the baby during a delivery..."

It is safer to never go on a boat. It is safer to avoid travel. It is safer to never drink more than a couple bottles of beer. Safety is not a goal to be achieved to the exclusion of all else. Of course we want the mother and child to survive, but in many cases medical interventions do nothing to improve survival rates.

Hospital births are unpleasant. The medical staff wants to get a live baby out. It matters little to them whether the mother has a 3 month recovery or a 4-week recovery. They don't care if you go home exhausted and stressed out. Many OBs seem to see their patients as unruly children who must be told what to do. They often seek control and predictability where a less predictable and more organic birth would be better for everyone involved.

Pitocin, epidurals, Caesarean sections, and other interventions all have legitimate and justified applications in some pregnancies. Like much of American health care, however, pregnancy and birth suffer from an excess of medical intervention. OBs certainly have their reasons. Our society is litigious. The financial incentives are perverse, rewarding the amount of work regardless of appropriateness. The staff doesn't have to suffer through a recovery made excessively difficult by unnecessary interventions. That their behavior is understandable doesn't mean it's in the best interests of all families.

Medicine is, ideally, an empirical discipline. However, the AMA cites no medical studies in support of their statement. That makes sense, because those research studies don't exist. There are studies suggesting that home births are not worse in terms of mortality, while being superior in reducing medical interventions. I am not the AMA, so I can't say what their motives are. However, the AMA is a powerful lobby with many characteristics of a guild. It seems reasonable that their motivation is to maintain control of a significant health area. That preserves their prestige as well as their livelihoods. Their efforts have the effect of reducing competition. That is so obvious a consequence that it cannot be accidental. Perhaps I am too cynical, and they make these efforts only with reluctance. Regardless, they consider the costs acceptable, which is suspect because they bear few costs and yield only benefits.

The final and most egregious part of this statement comes at the end:

RESOLVED, That our AMA develop model legislation in support of the concept that the safest setting for labor, delivery, and the immediate post-partum period is in the hospital, or a birthing center within a hospital complex...

They want to use the law to restrict individual freedom and force their methods on everyone. Their methods are often what's best. It's not often enough. The best data we have are clear, and the AMA offers little in rebuttal. Rather than prove the superiority of their care, and rely on individuals to make the right decision, they would rather use their prestige and political power to try to eliminate alternatives. That frees them of the burden of demonstrating their superiority, as well as eliminating much of the incentive to improve what is clearly not good enough.

This, among other issues, has led me to conclude that part of the dysfunction of the American medical system can be blamed on the AMA and similar organizations. They can be truly excellent in a number of areas, but they seem to believe that their expertise is broader than it is. Everyone make mistakes, even with the best of intentions. However, the AMA is a political entity as well, and those politics have tainted what they do. We have excessive respect for doctors in our society, and that reverence is hurting us all.

MS CS

noreply@blogger.com (Ketan) — Mon, 21 Jul 2008 14:57:00 +0000

Every now and then I toy around with the idea of getting a Master's degree in Computer Science. There's a lot to recommend it. I'm the least educated person in my family. My mom has two bachelor's degrees, my dad a BS and MS, my sister two bachelor's degrees and an MBA, and Jessica a BS and MA. I just have my lowly BA. It's not even a BS even though it's CS. Rice didn't offer the BS until my junior year, at which point changing course would have required at least another semester, which, at $10,000 a pop, was a bridge too far.

There are a number of important topics in computing where either my initial degree was weak, or the last 8 years has allowed my earlier knowledge to atrophy. Compilers, for instance. It's a bit of a chicken-or-the-egg problem with learning those things on the job, as jobs that use those skills require you to already know them. To some extent I can learn those on my own, but I would benefit from a more structured program for some of the more challenging and/or abstract subjects.

UT has a Software Engineering Master's degree program. I went to an information session for that a few years ago, and concluded that I'd rather have a proper MS CS. The SE Master's is a lot more vocationally-oriented, and the covered material either things I've already learned or could more easily learn on my own than the harder-core MS CS material.

On the other hand, there are lots of reasons not to attempt the degree. The most significant is the opportunity cost. It would take at least 3 years and $35,000 (at 6 credits/semester), at a time when my presence at home is pretty important. I don't think it would grant me a whole lot of earning power directly, though I should not discount the opportunities made available to me by knowing what I previously did not.

I would definitely like to have the knowledge, but it's not just about what I want anymore. My time and money are limited and already spoken for. It will probably be easier in 5 years, but it won't be as valuable then. That's one of the things I've realized. Getting older means that the opportunity costs of changing direction go up and the benefits go down. When you're young, opportunities multiply as time passes. At some point around age 20-25, things change. Now the passage of time means more doors close than open. Maybe it reverses yet again later when the kids get into (or leave) college, but that's a long time from now.

"The Dark Knight" review missed opportunity

noreply@blogger.com (Ketan) — Fri, 18 Jul 2008 23:07:00 +0000

The review headline in the "Austin-American Statesman" for "The Dark Knight" called it a "Fierce Tour de Force." Hello? How about a "Tour de Fierce?"

Drying clothes in a hurry

noreply@blogger.com (Ketan) — Tue, 15 Jul 2008 01:25:00 +0000

Use the microwave carefully to dry small articles of clothing if you're in a rush. Do it in 30 to 45 second phases. Doing it in bigger chunks is risky because if your clothes get dry partway through, they'll char and burn. Not fun. Don't put anything with metal in, either.

If you're not quite so daring, or your clothing is too big, you can use your regular clothes dryer, though it takes a little longer. Don't just put that one item in by itself, though. If you do that, it'll just stick to the drum by centrifugal force; it won't tumble. Thus, it won't get exposed to the stream of hot air. Put a couple of bath towels in with it. They'll knock it around. As a bonus, they'll also pick up some of the moisture. And, of course, if it's something you need to iron anyway, you don't need to dry it all the way.

A classic example of my law of ranking

noreply@blogger.com (Ketan) — Mon, 14 Jul 2008 18:37:00 +0000

Remember my law of ranking? From a blog post about the relative value of professional sports franchies in the US:

The New York Yankees are the only non-NFL franchise in the top 27.

Emphasis mine.

Why 27? Simple: #28 is the New York Mets, another non-NFL franchise. What sounds more impressive, the above statement, "1 of the top 25," or "2 of the top 30?" 27 is slightly bigger than 25, so presumably has slightly more impact*, and 2 is much bigger than 1, so its impact is much less.

* Not worth it in my opinion, especially considering the Law of Ranking, but whatever.