Amit's Thoughts

From Amit’s Thoughts

Heroes Season 2: Evil and dumb #

It seems that many of the main characters in Heroes, Season 2, are becoming either evil or dumb, and they're coming in pairs. Last season Jessica was evil and Niki as dumb. Let's go through the list for season 2:

• Kensei is becoming evil, and Hiro is becoming dumb.
• Nathan might be evil, and Matt is dumb.
• West seems to be somewhat evil (and somewhat dumb), and Claire is really dumb.
• Noah has returned to evil, and The Haitian I'm not sure about.
• Elle is evil, but has no dumb counterpart (yet); maybe she'll pair up with Peter, who's not really dumb, but has lost his memory.
• Maya is turning to evil, and Alejandro I'm not sure about. Or maybe Sylar is the evil one and Maya is dumb.
• Niki/Jessica has turned into evil Jessica, and she's working with Mohinder, who's dumb.

What if Niki/Jessica was the first to experience some sort of evil/dumb pair virus that will spread throughout the hero community? I know, it's a ridiculous theory. I can't even fit Angela, Molly, Micah, Monica, or the mysterious Adam into this list. But every time I see one of the above characters it does seem to me that they're getting more evil or more dumb.

I'm also starting to think Bob may be the only good hero. But then, I'm probably the only one who thought Linderman was good. He was the one who wanted to save the world (from losing 93% of the population) by setting off a small bomb (losing 0.07% of the population). Evil means but a good end. Except he's dead now, the bomb didn't go off, and now the world is in trouble.

Labels: heroes

Feeds are not enough #

There are lots of types of information I want to search and browse, but it's typically information that I want to see right now. I also want to be notified about new information and changes to information. Emails, text messages, and pager alerts are ways to send this information to me. Feeds are a little different in that they're controlled by the receiver pulling information rather than the sender pushing it. All of these systems allow senders to transmit messages to receivers. However all of them treat each message separately.

What I really want is information, not a set of messages. The messages should be grouped and summarized. Gmail for example groups messages into conversations. Facebook groups and summarizes friend updates. For example instead of separately telling me, “A is a friend of D”, “B is a friend of D”, and “C is a friend of D”, Facebook will tell me “A, B, and C are friends of D”. They can do this because they know the structure, not only text, of messages, and also because they know when I last read messages.

There are lots more things I'd like to see along these lines. For example if I receive 100 messages telling me my site is down, and I'm away from my computer, I'd like them to be combined together. Or if the site is back up maybe those messages should go away, replaced by a note saying the site went down and back up. If I get a traffic alert it should expire when the traffic clears up again. If I read news every hour I want to see what's new in the past hour but if I read news once a week I want to see the week's biggest stories, not the 168 hours of updates. When I come back from vacation I shouldn't have lots of low-importance and redundant messages. Group and summarize. Show me what's important. Don't overwhelm me with every individual message.

Labels: future, structure

Online map sites #

I love maps. I especially love online maps. They let me zoom, pan, and change the features being displayed. Google Maps was truly wonderful after using Mapquest, etc., for so many years. Microsoft and Yahoo also have draggable maps now.

I was planning a trip to Mount Saint Helens recently and tried out some online maps:

• Google:
• Yahoo:
• Microsoft (Live):

There are three things I really want to know when visiting Mount Saint Helens:

1. Roads. I want to know how to get there. Only Google shows the roads clearly. Yahoo shows dark gray on medium green, which is barely visible. Microsoft shows dark gray on medium gray, which is even worse, and nearly impossible to find unless you already know where to look. Just try to find the western end of NF-99 on Microsoft's maps. None of the services offered as much information as the park maps.
2. Terrain. I want to know where the volcano is and where the crater is. Only Microsoft shows this at all, and it shows it beautifully. I can see the ravines and mountains and river valleys. It's great! Yahoo and Google show nothing about terrain in their maps. Instead, you have to switch to Satellite view, which works for this volcano, but no match for the clarity of Microsoft's maps. (I'm sure terrain slows down Microsoft map loading quite a bit though; maybe it should be an optional layer.)
3. Stops. I want to know where I can stop, take pictures, go on a hike, etc. None of the three mapping sites I tried have this information; I instead got some of it from the park map. Another way to get this data would be from geotagged images; if I could see the most popular spots for photos, I'd know where I should go.

The other thing you might notice from these three maps is that Microsoft's typography is excellent. Google has a nice font but surrounds the text with bright yellow or white to increase contrast, which is quite distracting. Yahoo's text is decent and has less distracting contrast. But Microsoft's text is really nice. Instead of a white border they use a subtle white shading to increase contrast. It's very clean. They also use different fonts for different types of features — compare “WASHINGTON”, “Mount St Helens National Volcanic Monument”, “Spirit Lake”, and “Mount St Helens 8365 ft”. It looks more like a “traditional” map than Google or Yahoo.

In the end I used the park map for roads and features, Google maps for driving times and directions, Microsoft maps for getting a general sense of where the scenic areas are, and my Rand McNally U.S. Atlas (paper, not online) to find out which routes were marked “scenic”.

I also visited the Seattle area, and compared the maps:

• Google:
• Yahoo:
• Microsoft:

In urban areas Microsoft continues to have good typography with labels that are easier to read than Google's or Yahoo's. They also seem to be placing the labels more intelligently to avoid drawing over important areas. Microsoft also has labels for more items on the map. The terrain doesn't play a role here, although the east side of Mercer island has terrain that explains why the roads curve so much. Google's roads are easiest to see. Yahoo shows each city in its own color, which can be useful at times, and it also makes highway entrances easy to see. Microsoft shows more roads (faintly), which gives me a better sense of which areas are sparse/industrial and which areas are dense/residential.

Which site do I use most often?

In general I get more information out of Microsoft's maps than Google's or Yahoo's, but Microsoft has a few annoying implementation details that keep me from using their maps more: (a) their awful browser detection script rejects me unless I lie about my user agent, and (b) I get a stupid dialog box asking me to install a Windows 3d plugin … even though I'm not wanting 3d, nor am I using Windows. There are things about each of the three that I like, but in the end I'm still using Google's maps most often. It's fast and the roads are easy to see, especially in cities.

Update: [2007-11-30] Google now has a “terrain” map mode that gives me what I liked from Microsoft Maps:

Go to Google Maps

Labels: review, typography

The Price of Gas #

Prices are typically driven by supply and demand. I was curious about the price of gasoline. When I buy a gallon of gasoline, I pay for it, but others pay for it too. My purchase increases the aggregate demand. Higher demand means higher prices. Higher prices means other people pay more for gas.

How much more do others have to pay?

I can't calculate exactly but with some simplifying assumptions I can make an estimate:

1. Limit this calculation to the United States. There are complex issues that influence gas prices around the world, some economic and some political, and it's much simpler to make this estimate with just one country. This would ordinarily not work, except the next assumption makes it possible:
2. Gasoline supply in the U.S. is limited by refining capacity in the U.S. With “maintenance”, “shutdowns”, “inspections”, fires, and other capacity issues, I believe it's reasonable to say the supply — at least in the short term — is fixed, and that everything that is produced is consumed.

In other words, if I buy one extra gallon of gas, other Americans need to buy one less gallon. What would it take to make them buy less? Raise the price. By how much?

Let's call the total quantity consumed by everyone else Q. Let's call the current price P. We want to know how much the price has to go up to make Q go down by 1. If everyone else (collectively) buys 1 gallon les, then I can buy that gallon. The key to the relationship between P and Q is the price elasticity of demand:

e = %ΔQ / %ΔP = (ΔQ/Q)/(ΔP/P)

What I really want to know is when I spend $D on gas, how much more do other Americans have to spend? Spending $D means ΔQ = D/P, which can also be written as D = ΔQ×P. What everyone else has to spend is Q×ΔP. So let's compute Q×ΔP. First, let's rearrange e:

e = (ΔQ/Q)/(ΔP/P) = (ΔQ×P) / (Q×ΔP)

So Q×ΔP = (ΔQ×P)/e = D/e.

When I spend an extra $D on gas, others have to spend an extra $D/e on gas. That's the answer I was looking for.

Except … what's the value of e?

There are various estimates: 0.2, 0.01, 0.1, 0.034 to 0.077 in 2001-2006, and 0.26. When I spend an extra $40 on gas, other Americans have to spend between $153 and $4000. I'm not sure which to believe, but I'm going to guess it's around 0.1, which means others have to spend an extra $400 on gas. Where does that money go? To the oil companies.

Let's look at it in reverse: if you found a way to spend $40 less on gas (maybe carpooling, planning errands better, or driving less aggressively), not only would you save that $40, the oil companies would miss out on $400 (maybe as much as $4000), because you'd be helping other Americans spend less on gas.

I'm not even going to try estimating how much more everyone pays when someone drives a big SUV instead of a fuel efficient car…

Labels: economics

Dividing a range into segments #

Suppose you want to divide a numeric range (such as 0–1 or 0–23 or 1–365) into even segments. If you know how many segments you have, it's easy; you divide by N. But if you don't know how many segments you will have, and you can't go back once you've divided something, it gets trickier. If you divide into 3 equal segments and need 3, you're at the optimal point. But if you instead need 4 and have already divided into 3 segments, you end up subdividing one segment of length 1/3 into 2, leaving you with 4 segments of length 1/6, 1/6, 1/3, and 1/3.

There's a clever division scheme involving the golden ratio:

1. Rescale your range to be from 0–1.
2. The i^th division occurs at i * φ

It's so simple. Why does this work? I don't know. But it's pretty neat.

I first ran across this when I was looking for a way to pick sample points in 1 year of data. I wanted a set that would be roughly evenly spaced, because I wanted to draw a timeseries chart with the results, but I didn't know how much time it would take to analyze the points. So I analyzed one at a time, using the golden ratio to guide me.