Monday, September 27, 2010

Moved

Quick shoutout: I've moved my learning blog to Tumblr, at metacog.tumblr.com. This blog is frozen for posterity.

Thursday, August 5, 2010

Google Kills Wave

Half a year ago I wrote:
Is all lost? The answer to this isn't clear. It could be that Google Wave would, against all logic, live up to its hype. It could also be possible that Wave flops, completely, and is discarded as Rasmussen and co. leave the project to work on something (hopefully) more useful. Or it could even be possible that Wave turns out to be a sleeper hit - that it fails in its intended purpose as a communication platform and is then adapted for uses its creators never thought imaginable.
I had no idea how right I was.

Monday, April 19, 2010

Don't Be Strangers

I joined CS3216 because I was looking for a good simulation of a startup environment.

I think much of my school life has been about simulations. I created a debate club because I wanted to simulate leadership - especially against rabid school opposition; I played Judo because it was a good simulation for pain. I'm not sure if they're perfect simulations, but I try. And so - for my application to 3216 I wrote:
This is probably the only time in my life where I may learn the realities of a startup without the threat of monetary failure, and I must admit that I was saddened when you told us that this would be the last 3216 offered in NUS. I originally planned to wait a year before applying; I believed I would’ve been a better programmer then.
But I'll say it: CS3216 has been an amazing simulation.
 
I am supposed to tell you about what I've learned from CS3216, but I'm not sure how to do that without sounding corny, or staid, or weird. So ... let me tell you a couple of stories.

Networking should be immediate
The night after the entrepreneurship lecture I hunkered down at PL6 to take a stab at Sportup's design. (It was a wasted night - nobody liked the design and so I had to start from scratch a couple of days later ... but anyway). Sometime after 11 Tomithy came in, bringing food with him, and he sat down with a whole stack of name cards in front of him. "What are you doing?" I asked.

"Emailing." he answered.

And so he did: I glanced over now and then, and he was shooting off emails to the people he'd met at the entrepreneurship session. I thought to myself: now this is something I've got to learn to do.

But of course I didn't email the people I talked to that night, and of course I didn't get anywhere on that. This must change, of course, and soon. (An aside: Toms is amazing. Did you guys see how brilliantly Supremacy was executed?)

Chemistry is difficult
I have to be honest here: I made a couple of big mistakes in CS3216.

The first, and biggest, was to choose the wrong teammate. Now I'm not sure if wrong really means ... 'wrong'. I believe 'wrong' really means 'wrong for me', or 'wrong for us.' It could be that a hundred and one things got in the way, and I just didn't make the effort to make sure we were on the same page. But if I'm honest with myself: I think now that we shouldn't have worked on the same projects. We'd have saved each other a lot of heartache that way.

I have learnt that chemistry matters - and it can be impossible, sometimes, if you've got the wrong mixture. Our personalities were direct opposites, and we came from completely different world-views. I think, on one level, I should've stepped back and let him find a better, more suitable teammate. But on another, it's good that I've learned this now - in the simulated environment that is 3216; better here than in the real world, with an actual startup.

Some relationships just won't work out - the trick is to figure out which.

Things came to a head between the two of us, the night leading up to the final project presentation. We were in COM1, and we were ... loud in our disagreements with each other. So I went downstairs, to vent to Orry and Hung. And the advice they gave - they were amazing.

"This isn't a problem about design," Orry said, "It's about something else. Something bigger. You've got to figure that out before you attempt to fix it -"

"But there's no time!" I said.

"Yes, there's no time." Orry said, and there was pity in his voice. I slumped in my seat.

I'm not sure what I could have done differently. Working on a project in CS3216 is a little like marriage - and if you pick the wrong partner, no matter how hard you work at it, you're doomed, rather, to fail. And that's a humbling thought.

I wish I could've gone back to change this - all I have now are apologies to my teammate, and that's not much.

Beware the Dunning-Kruger Effect
I got my first taste of the Dunning-Kruger effect this semester, and I'll keep that in mind the next time I work on another high-stakes project.

Learn to stay focused
This from Hung, on my teammate problem: "I think one thing is that your design isn't fast enough. Like - I can't worry about (Supremacy's) design - because I've got so many other things to do. But he could, because he was so far ahead than where you were. You were slow with the design. I think ... I think it's because you're not focused enough. You can get very distracted sometimes. This is not something you can change immediately. It takes time. Slowly, perhaps."

Hung is right.

Derive arguments from first principles
Prof actually attacked me for some of my essays. Lesson learnt: derive arguments from first principles. I'm still trying to figure this out, because some arguments are too big, and too intricate that deriving them from first principles would take a whole book (e.g.: the argument that fairness is objective takes 100 pages of intricately argued logic), but this is, I think, a principle worth holding on to.

I am thankful for this - too many of my essays, when I look back on them, have jumps in logic that are clear to me, but not to an average reader.

An algorithm for taking criticism
Having your professor attacking you for your own good was tough, but from that experience I think I've finally figured out how to take criticism, and it goes something like this:
Upon receiving criticism, determine if a) person means well (wants you/your app to improve) or b) person does not mean well.

If b) disregard criticism.

If a), filter for two kinds of information: valid points and invalid points. Invalid points are points where you know the critic is wrong - for instance, if he/she is assuming prior experience that is completely different from your past experience - ignore this. Do not get defensive. Nobody knows your personality as well as you do, and in the course of giving advice, the person is likely to get several things wrong about you. This is normal.

Focus on valid points. Ask for (impartial!) third party opinion, from people who know you well enough.

Absorb valid points.

Thank critic.
There are different kinds of motivation. Star Wars will never be the same again.
I've a feeling Prof really likes Star Wars. He thinks he is better looking than Yoda (though Yoda is kinda cute). Okay wait, that wasn't the point I was trying to make.

The point is this: motivations matter. For the longest time ever I was motivated by ... I don't know, really. Wanting to prove detractors wrong. Or wanting to prove myself to myself. Prof said this was bad. Much of his advice on the matter was shrouded in geeky metaphors like the 'dark side' and the 'way of the force', but he was making an important point. Motivations matter - and some are inherently better than others. I will be leaving CS3216 thinking about my motivations, and I'm not sure about the answers I may find.

I'll pray about that. And I'm fairly certain I'll find answers, sooner or later.

Doubt is normal
Doubt is okay. It is normal.

I stayed up nights wondering if I was on to something on Treehouse (I'm still not sure). I delayed launching Pandamian for a full year because I thought it was going to be disruptive to the publishing industry - too disruptive for a 20 year old to launch on his own. CS3216 has taught me that it's okay to doubt - and I learnt this from talking to Kah Hong, to Reuben (who himself was doubtful over Voucherous). It's cool to know that there are other people out there who also doubt, who also fear. I think - when Pandamian launches, one of the first things I'll have to do is to set up some kind of informal support system between founders - because the doubt, the stress, can really be too much for one person to bear.

But I'm not sure, of course. Too many things can happen, and there's no telling how things would be like even a couple months down the road. What I do know is this: if you're going to start a startup, do it with someone else. The doubt can be overpowering otherwise. That's a valuable lesson indeed.

And, lastly:
"Don't be strangers, when you see each other in school, in the future. I wish you the very best."

Don't be strangers
'Don't be strangers' was probably the last thing Prof said to us after the poster session. Looking back on the course, I now realize that the people I'm closest to are the ones with whom I've pulled all-nighters/spent long periods together with. Kah Hong has been a real blessing, in terms of support, and he's one of the few people in the class who has a real sense for 'what users want'. I've spent nights in COM1 with Orry, and Hung, talking about all sorts of weird things. But I think this connection applies to everyone in the class. There are too many other people with whom I can now nod at - in the hallways, the biz canteen - and know that we've shared a semester of pain together - and survived.

I entered CS3216 alone, and left with friends. And I can only be thankful for that.

Onward.

Monday, April 12, 2010

Liveblogging the Last Lecture

The class is full, and there are extra people attending the last lecture. Lots of people milling about, joking, doing very 3216-community things.

We start off with a short pitch by the guy behind Mobiwerkz. If you want to join, and you're interested, go email him.

7:00 Watched a video

7:05 Everyone jolly well go and write blog entry, to tell what the heck you learnt. Crowd sourced to score people our blogs. Clever bastard.

7:07 "So it's all your fault!" Big grin.

7:08 To Shannon's suggestion that everyone puts a 10 for everything: "But that won't work! Why? ... Exactly! Prisoners dilemma!"

7:12 (Blogger's note): Oh great. I have a word limit.

7:13 Remember "Wong Lim Soon!" And, yes, he's the prof that determines everything.

7:14 Sales matters

7:15 People matters. Whoever you actually work with matters. (Oh yes, definitely.)

7:16 Execution really matters. The execution has not been fantastic this semester. And there's a picture of a bunch of kids climbing a tree. So there's two things about execution: 1) try random stuff 2) keep trying. And then there's working on the problem, and listen to people. Listen to prof. Which is a balance between being stubborn and sticking to your principles, and listening to people giving advice.

7:17 So there are many paths to success.

7:20 Anybody want to go back? Got a hot date to go home? (A couple seconds later) No? So I shall now indulge in my favourite past-time: preaching to students.

7:23 I'm not a psychologist, period, but I try to observe things, and derive conclusions, and what I've seen is that nature sets the limit on what you can achieve. This is what ... I've seen.

The nurture part determines whether you can achieve it.

7:24 So the truth of the matter is none of you can be a Tiger Woods. The golf arr, not the sex! /uproarious laughter.

7:25 There are two things. There's nature, and there's nurture. And sometimes life's not fair.

Some people complain too much. The good news is that most people's capacities to achieve is much much higher than what they actually achieve.

So that gives me a job.

7:28 Attitude. Ultimately, what I hope to do in this class is to fix mindsets. (Insert section about Jolly Good Fellow engineer in Google, and talking to them, and finding out how to think.)

7:32 A lot of people are poor because their minds can only handle so much money. They are programmed to be poor. Quote from Secrets of the Millionaire Mind. You agree, don't agree?

7:34 Yanjie disagrees. Toms says that it's not particularly concrete. Prof steps in: "Nature nurture rubbish again."

7:36 Generally, however, Prof thinks this is true.

7:37 "Give a poor guy 100k, and at the end of 10 months he was a lot worse than when he started off."

7:39 How you perceive the world really matters. Research has shown that optimists are more likely to succeed than pessimists by a long shot.

7:41 I believe education has three things. 1. Find direction, 2. Fulfill their potential 3. Become good people

This whole thing, however, has this big 'It Depends'.

7:45 Confucius says: self, family, country, world. There are these things that he says. Then he says, you figure out how to govern yourself, then you figure out how to help your family, then you figure out how to serve your country, then you 'attempt to bring peace to the world.' To change the world.

Self
7:47 Learn to like yourself. To be at peace at yourself. To be okay with the way you're born. There are people with some deep-seated insecurities. Know what you're capable of, what you're good at, focus your strengths, and outsource your weaknesses.

4:58 Delegate - you can delegate responsibility, but you can never delegate accountability.

4:49 Deal with fatal flaws. And the worst fatal flaw is laziness.

[Though there's this military quote: "I divide officers into four classes -- the clever, the lazy, the stupid and the industrious. Each officer possesses at least two of these qualities. Those who are clever and industrious are fitted for the high staff appointments. Use can be made of those who are stupid and lazy. The man who is clever and lazy is fit for the very highest commands. He has the temperament and the requisite nerves to deal with all situations. But whoever is stupid and industrious must be removed immediately." - Attributed, circa 1933; General Baron Kurt von Hammerstein-Equord (1878-1943); German Chief of Army Command (1930-33)]

7:50 Take personal responsibility (or - jokingly - find a religion and blame God).

Failures come from two sources. One is it's your own fault. And the second one is ... act of God. That's life. Bo pien.

7:52 Stay teachable. Try to keep learning. And try to learn new things. Prof praises Reuben and Ryan, for posting new things, because they keep reading new things. They read, he reads.

7:54 You're less important than you think. Do not to be too concerned about how you look. People won't remember you.

7:56 Actually I have a problem now. Now I've taught for four years, and I can still remember all of you. But in one or two years, I will have to start to garbage collect.

7:57 10,000 hours - it turns out that I've done 10,000 hours of teaching. But I don't want to. 20,000 hours.

7:58 The char kuay tiaw experiment: go out one day and cook, without my skills, to see if it's possible to make a living without an education. And I think I can make more money than as a prof.

7:59 I am better than I was yesterday. Quoted for truth, hopefully dy/dx > 0; but nature limits you. But remember that the limit is an asymptote. You can't reach it - so as long as you keep getting better, it will be okay.

8:00 There is no substitute for hard work.

"So part of the thing to do as a lecturer is to make sure nobody dare to sleep"

Hehehe.

Family
8:01 If you marry the wrong person, you. Are. Screwed. Photo of Ris Low appears on screen. /laughter

8:03 "So what qualities are there for a spouse?" Pause. "From experience?" *Laughter*

Prof looks at us. "It depends."

8:04 The girl that's right for him (points to Reuben) is probably not the girl who's right for him (points to Adhiraj). Sorry I sounded like a scientist, because I am.

8:05 For the first few months of being in love, there's a chemical in your brain called PEA that makes you high - typically lasts six months to a year. Then it wears out. So the moral of the story is: don't marry someone in six months. There are people who marry and get it right ... but who feels lucky today?

8:06 Right time to marry is when she 'feels like family'. So there's a probability that she won't be sexy forever. One kid still okay, two kids a bit cannot make. /uproar and then pause. Oh you mean you recorded me ahh? Die lorr.

8:07 Be careful about these things, because it's not so simple. For Christians, look for the Proverbs 31 girl. And, err, girls don't look for the Proverbs 31 guy. Because otherwise he doesn't do anything.

8:10 So there's this story about a guy walking through a field, and he was asked to pick the tallest head of corn. And he came out empty-handed. Then he was brought to a forest, and told to chop down a tree and bring it our (that is smallest and can be brought out) and he came out with this puny little thing. And that is love.

8:11 Not easy wan la. If you can find an algorithm to solve this, you'll be very rich (Shannon behind me says: there's this episode of How I Met Your Mother ...)

8:12 Don't do a Jack Neo. (Class laughs). So that brings me back to Character vs Reputations. Typically character implies reputation. But reputation is not if and only if. So people with good reputation may not mean good character.

Country
8:14 Skip the government bit. Because it's Singapore.

The World
8:14 Think. Now here are a couple of articles that I've taken from the Strait Times.

/shows an article from ST about this guy who is a top chef. So this guy is a gifted chef. And he probably makes money than you do. The thing is, do what you're good at. If this guy tries to be a prof, he will sure die.

8:17 Shows a picture of Thailand. They're a failed state now. A bit sad. But why are they fighting? Class says it's red shirts vs yellow shirts. So Singapore doesn't allow protests. And it's complicated. They're arguing about human rights, but they may not know what they're talking about. So they have to think. Think. Things are not so simple.

8:18 Shows a picture of this 50 year old guy, who's unemployed, headlines: where are the missing fifties? And the times are changing - during your grandfathers time, when you have a degree it's a big thing, you're set for life. And when you get hired. you'll be taken care of by the company.

So what has changed?

8:20 Too many people have degrees. So what has changed hasn't been about the degree. If you're useful, you'll get hired. If not, you have pretty degrees, but you'll 'die die never get hired'.

8:24 In order not to loose your job at 60, just be damn useful. That is the best thing you can do.

8:25 A few words on the stock market. So the stock market has been rising. And this has been happening since the industrial revolution because the people have been putting their retirement money into the stock market.

So are stock markets going to be rising all the time?

Not sure. You can't base it on past performance. But there's going to be huge demographic changes. Where the bottom constricts. And the top (old people)expands.

8:27 If it's too good to be true, it probably is. Small smiley face with a label: "Ponzi Schemes."

8:28 You cannot achieve superior outcomes in education without the teachers. So I realize something that I haven't done so in the past, and that is to encourage my students to become a teacher.

The thing about teaching is that it doesn't play very well. But it does pay okay in Singapore. So consider it.

8:29 I can't tell you what's going to happen in the future. With 3216, that is. So there's going to be two courses. And semester 1 will be 3215. Which is SoC only, basically to train up the software engineering part. And for semester 2, we'll absorb all the non technical people. So most of the technical learning would be done. If any of you have good ideas on how this is to be done, please contact me.

8:31 Long term goal is to produce 60 students of this calibre. And ... I don't really know how it'll be.

8:32 So if you can't remember a thing from this lecture, take this: figure out what you like, AND you are good at ... and just f*cking do it. Because there has to be a balance between liking it and being good at it.

8:34 Do the right thing. I think in life, there will come a point in time where there are two options. One option is to do the more profitable thing, and the other option is to do the right thing. Sometimes you don't know. And if you don't know, and you just do, it's okay. But if you know there is a choice ... then I want to encourage you to do the right thing.

And doing the right thing usually has a cost. And sometimes you don't even feel shiok. Most times, in fact, you will pay. But ... doing the right thing typically, as Randy Pausch says: the 'karma will take care of itself.' And while that's a very buddhist thing, it's the truth.

8:36 I'd like to thank my teaching staff. /applause

8:37 ... and this is for you. *shows slide with image of book titled 'Common Sense'* There are many problems in the world, and a lot of people don't have common sense. And I think a lot of it has to do with ... being curious about the world. To gain common sense, that is.

8:38 And so, with that, I wish you all the very best for Wednesday. And this is the end ... but it's also the beginning. Thank you very much.

Thursday, April 8, 2010

How To Do UI Design, In 8 Easy Steps

When Prof Ben asked us what workshop we think should've been included, in the mid-term feedback, I remember leaving the field blank because I couldn't think of anything to write. I now regret that decision. In the past couple of days I have come to realize that people don't understand the underlying mechanics of design. I have been embroiled in a two-day-long argument with what to do about Sportup's look, and I think it would do to explain some of the most basic principles here, on this blog, once and for all. I hope this would be of some help to you; if it doesn't make sense - feel free to shoot me an email, or a comment if you're so inclined.

What Is Design?
Design is how your users interact with your product. That's all there is to it. Simple as pie.

Now, there are two misconceptions about design that I want to clear up immediately: hopefully for the first and last time. The first misconception is that good design has to be pretty. The second misconception is that good designers have an innate feel for beauty, and so therefore design by sitting in front of the computers and working until something feels just right. It does not. They don't. Design isn't some wishy-washy thing that people with easels and paintbrushes tucked behind their ears waste their time indulging in. It is as much a science as it is an art, and with computers - the field of study that is related to design is called Human Computer Interaction. There are principles in HCI, just as there are principles in software engineering, and the results for breaking them are the same in both: you end up with an unusable app. What are they? I'll come to that in a bit. But before I do so, let's talk a little about the two misconceptions:

Misconception 1: Good design must look pretty
You may not believe me on this, so let me show you a couple of things, to make you understand.

The Drudge Report broke the news on Bill Clinton's affair, and till this day continues to exist with the exact same design it had back when it was first created. Jason Fried of 37signals calls it one of the best designs he has ever seen, and wrote a 1000 word article on why he believed it to be so.

Jakob Nielsen is the world's leading authority on web usability and design. His site looks like a piece of shit. But when you're there, you find something odd happening: you know immediately where to go, what to read, and where you are on the site at any point in time. Nielsen's site has good design.

In 2005, Blogger contracted Douglas Bowman of Stopdesign to create and implement a host of themes that are still in use today. Of those designs, Bowman created one that is almost universally lauded as timeless: a simple, bland, minimalistic theme he named Minima, the same theme I'm currently using on this blog. Why is Minima considered good design? Simple: it's used by hundreds of Blogger blogs, all over the world, for all kinds of topics, by all kinds of people. It is timeless. Ugly? Yes. But good design all the same.

Still not convinced? Well, consider this: when Paul Graham built Hacker News, he purposely chose a garish colour and a sparse, ugly design. He made the site for programmers to use, and he wanted to discourage all other kinds of people from joining the site (and therefore the discussions). Hacker News is today one of the few online communities with a consistently high level of intelligent commentary, and this is by no means an accident: Graham's built all sorts of clever design hacks into it to encourage just this level of discourse. Ugly it may be, but bad design it is not - HN's look does its job, and does its job well.

Now you're probably going to say - sure, dude, use bad design when you want to attract programmers (who don't have a sense of design anyway) ... but when you're building for the real world you have to look pretty! Really? Wrong. Two of the most successful sites ever built on the Internet have been based on particularly weak aesthetic foundations, and if you don't believe me, take a long, hard look at a) the Google home page, and b) Craigslist.

Misconception 2: Designers design by feeling.
They do not. I'm not going to lie and tell you that designers are as hard-assed about logic and testing as programmers are, but on the other hand they're not some sculptor who plays with clay and goes through a thousand iterations before they finally find the inspiration to make 'the one'.

If you're working with a good designer, anyone who's done some outside work before, you'd know this. When I was in 9rules I complained to Mike Rundle, the designer-in-chief in charge of the site's overall look and feel, and he told me to wait awhile, to see if things changed. And it did. Mike's original design didn't feel right, but over time I grew to love the way the site worked, the way it felt. The 9rules community is, to this day, one of the best things that has ever happened to me online.

Designers who know their stuff work around several universal principles. When I visit sites with friends, I often hear them complain about this design or that design, saying things like: "Oh that's just ugly. How did she get so successful?" or "I think I can design better than that."

What they don't get is the simple truth: they can't. And why? Well, that's simple too - they're missing the point. Design isn't about aesthetics. It isn't even about choosing the right colours. Design is - at its very core - about building the right things the right way, and doing it in a manner that is befitting of the user.

Everything else is secondary. Even looking pretty is secondary. A user who can't use your site isn't a user at all - he's just a statistic. Here and gone in the blink of an eye. I should know - I've gotten plenty of those in the four years I've run my sites.

Eight Simple Steps To Good (Web) Design
As the title says, here are eight things that are pretty much universally true. I hope they do you some good.

1. Get their attention. Period.
This isn't true for most of HCI (which is concerned with the design and implementation of computer interfaces) but it is true of the web.

Internet users are fickle-minded people. Most visitors you'll get to your site will stay around for no more than 3 seconds, tops. The trick is to get them to pause - even for a little while.

Everything you'll learn about webdesign is built around the central philosophy of getting. People. To. Stick Around. For Awhile.

Get that into your heads. I spent four years figuring these things out, I'd only wish I'd learnt it earlier.

2. Important things above the fold.
The fold is what you see when you load a web page, and you don't scroll down. (Anything you have to scroll down to see isn't above the fold.

The term originated when newspapers still existed were first introduced. The fold used to be the top part of the front page - the bit that shows when it's folded in half and on a newsstand. Ads above the fold sold for twice the amount of money that of ads below the fold, and for good reason - those ads (and the headlines) were the first thing a prospective newspaper reader sees.

3. Keep to the left.
There is irrefutable evidence that users pay more attention to elements placed on the left side of a webpage as opposed to the right side. In fact, the actual figures are 69% (of attention) to the left, and 30% to the right. When designing a web page, place the most important elements to the left of the page. This is one reason why so many sites place their copy to the left - it gets more attention, and therefore makes it easier for the reader to understand what the site's all about.

Keep to the left for important content, you'll have far less trouble otherwise.

4. Design to the F
Screen shot 2010-04-08 at 1.17.38 AM.png

A direct implication of steps 2 and 3 is that the most important content should be placed on the 'F' of attention.

Users view websites according to an 'F'. Design in anticipation of the 'F' and you'll find that - oddly enough - users would know where to go and how to find things. Sites that employ the 'F' to their advantage are often described as intuitive, simple, easy-to-use. Use Facebook? Take a look at the current redesign. Where are the most important functions located?

To the left, and to the top, correct? That is the 'F' of attention right there, in action, acting right under our noses.

5. Understand visual importance
Visual importance is a little hard to understand. One of the simplest examples is that of colour. If you've got something small, and placed to the very right of the page (far out from the 'F' of attention) how are you supposed to draw attention to it?

One simple way is to use colour to your advantage. Colour it in with a brighter colour and you'll compensate for the element's position on the page.

Visual importance gets a lot more complex than this, though, and a lot of it is subtle stuff like rounded borders and smooth gradients (and the kinds of emotional responses they get out of you). If you've read this far, and you're interested in becoming a good designer, I'd recommend Andy Rutledge as a starting point. (Psst: read his post on design gestalts).

6. Keep things lean.
Choose a good width for your content. Wider content is harder to read. The optimal width, I have found, is 500-600 pixels. Any wider and you'll have to adjust by increasing your font size.

Don't believe me? Grab a nearby book. Chances are that there aren't more than 20 words on a single line. There's a reason for that - it turns out that a reader's attention tends to drift away the longer and denser the line is. I can't remember where I first learned this - but it's a principle I apply to all the writers who come to Novelr for design advice. Want to use long text widths? Then increase your font size - you'll have less words per line than you would if you didn't, otherwise.

(For the curious types: here's an article on the phenomenon.)

7. Use less columns
This depends largely on whether you're interested in being read. I'll admit here that this doesn't apply to most webapps, but if you're designing a text-heavy interface (such as a blog, or an ... Argo,) you'd do well to control the number of columns that appear on any one page.

User attention is split by number of columns. The more equal-width columns, the more distracted the user becomes. The solution is to either cut down on the number of columns your site employs, or to subtly emphasize relative importance using visual cues like shape and colour. How? Don't look at me - go out and experiment for yourselves. And that, perhaps, leads me to 8:

8. Test a lot. Use good tools.
... because that's the only way to know if the design's just right for you. Good tools, you ask? I personally enjoy Clicktale, and Crazy Egg. Though analytics and plain vanilla user testing works just as well too.

Good design is timeless design, and timeless design takes time (pun intended). I've only done one so far, and I'm still largely hit-and-miss with my work.

But I intend to get better at it. And if you're working with a designer who knows his stuff? Please, please stay out of his hair. He knows what he's doing, and feedback is good; but please - try to be fair.