Citibank is one of the biggest banks in the world. You would think they would have high standards in usability testing. Apparently not. In the last 5 minutes I have found 2 serious bugs in their web portal.

1. I have 2 messages waiting from them. I click the message drop down box and click the second message. Unfortunately it instead shows me the first message. In order to get to the second message I have to hunt around for the button to “list all messages” on a different page, and then click on the second message.
2. I tried to respond to a Citibank employee who was asking me what kind of checks I wanted. I wrote my response, but when I hit “send” it said “Can’t send the message. Don’t use any special characters, such as ampersand, slash, hyphen, etc.”. Well, I wasn’t using any of the characters they explicitly mentioned, so I must have been using one that falls under “etc.” So, I take out the dollar sign. Still doesn’t work. I take out the colon. Still doesn’t work. I take out the parentheses. This time it doesn’t complain, but instead tells me that there was an error and my message wasn’t sent.

Lessons for Citibank to learn:

1. Test with more than 1 message before deploying to your tens of millions of users.
2. Sanitizing your database input makes more sense than putting silly restrictions on what characters your tens of millions of users can put in a message.
3. If you insist on having silly character restrictions on messages, at least tell your users clearly what those restrictions are.

Ray Kurzweil you are full of shit. I will bet you $10 against your DVD with the ST:TNG episode of “The Game” that this won’t happen.

“By the late 2020s, nanobots in our brain, that will get there noninvasively, through the capillaries, will create full-immersion virtual-reality environments from within the nervous system. So if you want to go into virtual reality the nanobots shut down the signals coming from your real senses and replace them with the signals that your brain would be receiving if you were actually in the virtual environment. So this will provide full-immersion virtual reality incorporating all of the senses.”

When the iPad came out, many people questioned whether it could succeed. There were good questions about cost, about people’s willingness to adopt a new form factor, and about the experience that could be supported by fairly minimal hardware specs. Then there were bad questions, about whether a tablet can duplicate the functionality of a desktop machine (answer: it can’t. it’s not supposed to), and why a new tablet would succeed when older tablets failed (answer: because it’s different, and designers learn from the past).

The iPad is clearly a success now, or at least it hasn’t failed, and the tablet wars are in full swing. The biggest recent entrant into the arena is RIM’s BlackBerry PlayBook. I think the PlayBook situation is really sad, because in my opinion it is doomed to fail. It is doomed to fail not because it isn’t good, in fact I don’t know if it’s any good. I think it is doomed to fail because it has been marketed with one of the worst campaigns I have ever seen. So, let’s compare the iPad marketing with the PlayBook marketing.

iPad 2:

If you watch the iPad ads you get a clear message: the iPad will change how you live. Your wife/husband will be happier. Your kids will be smarter and better behaved. Your employees will be more productive. The iPad is going to usher in a new utopia of friendship and collaboration. Is it bullshit? To a degree, certainly. The iPad can’t do all that. But the message is a compelling one. Apple’s device is going to make your life better, and provide you with an experience that is memorable, positive, and compelling in many ways.

PlayBook:

The PlayBook ads have a single message: it runs Flash.

What? Seriously? Apple has promised to change our lives, and you promise to support a media format? This is completely meaningless. How many people even know what Flash is? I actually do know what it is, and even I am totally unimpressed by this claim. My message to the BlackBerry people is: I want to know how your device is going to change the way I live my life. You have told me nothing about this. You have failed in your marketing.

RIM’s failure is a real shame, because it is quite possible that their device is great. Unfortunately nobody will ever know, because nobody cares enough about it to give it a chance. And RIM, that is your fault.

I’ve decided to start learning how to work on old cars. One of the first things I’ve done is to figure out how drum brakes work. Drum brakes, in my opinion, have several very flawed design features. The first thing I discovered when I realized that my left rear drum was seized up. Something happened internally that caused the drum pads to expand and lock into the drums themselves. Rolling the car caused terrific squeaking. Unfortunately, the only way to disassemble drum brakes is to take the drums off. This is impossible if they are seized. Just to clarify: if your drum brakes are seized (i..e broken), you can’t take them apart to fix them.

Of course there is a fix, but it is a hell of a fix. There is a tiny little hole in the backing plate of the brake. You can poke a little screwdriver in there, and manipulate the internals of the brakes in order to loosen them up so you can take the drum off. This is almost impossible to do. I spent about two hours lying on the ground, contorted, sticking a screwdriver in a hole that I couldn’t actually see, poking at a thing that I wasn’t entirely sure was the right thing. The thing you poke is a wheel that has to go around a certain number of times before the brakes loosen. Unfortunately you can’t really tell if it’s loosening as you do it. So, I spent about 2 hours not knowing if what I was doing was having any impact at all on the problem. Luckily it eventually opened up.

That is a poor design. So, how to fix it? Couldn’t the hole be a little bigger? Just big enough to shine a light in while my screwdriver is also there? Couldn’t the wheel have little notches of some sort on them that can catch the screwdriver a bit better than what is already there?

I guess it is too late for these design fixes anyway. Drum brakes have been largely replaced by disc brakes anyway.

Next project: Fuel line.

I have a love/hate relationship with things that plug into other things. Sometime you plug a thing in and it just feels so good, the snugness of the sides of the male as it engages with the female. Other times it is infuriating, with sloppy connectors that don’t fit properly, or put unreasonable demands on the user. Here are some of my favourite and least favourite connectors.

USB: Pretty good. A USB connector feels great when you plug it in, stays firmly, and generally requires no configuration. But WHY is it symmetrical? Half the time I plug it in backwards and have to reverse it. If you add up the time lost from hundreds of millions of people plugging in USB connectors incorrectly, you’d have enough time to initiate and complete a moon program. I won’t go into all the mini-micro variants. They are pretty universally mediocre.

Firewire-400: All the good of USB, with none of the bad. The asymmetrical plug removes all the ambiguity. The only possible downside is that it may be a little big for small devices such as laptops. Plus, you get the bonus that Firewire allows for much better channel sharing than USB. Too bad it hasn’t really caught on.

Firewire-800: What happened, Firewire? You went from head of the pack to near the back. Firewire-800 plugs aren’t nearly as secure as their 400 predecessors, plus we have the symmetry problem again. Boo on Firewire-800.

VGA: Whoever invented VGA connectors obviously wanted to inflict as much pain on the universe as possible. The thing doesn’t stay in unless you screw it in. To screw it in you rip your skin on these tiny little knobs. Then the little bolts on the case come off when you disconnect the thing. Plus they tend to be jammed close to other connectors, to you can’t even turn the knob properly. WORST CONNECTOR EVER.

HDMI: A million times better than VGA, and more functional. Feels good when inserted. I’ve heard audiophiles complain that their ridiculous heavy cables sometimes result in the connector becoming unseated. I’ve also had problems with the bandwidth of HDMI being insufficient for my super high res applications. But as a connector, I like it.

Those are the connectors I have strong opinions about.

I’ve lived most of my life in Vancouver, and over that time I’ve witnessed the evolution of public buses used here. One of the things that fascinates me the most is the technology that the buses use to allow users to exit the rear doors. Interestingly, different buses have radically different approaches to support this.

As far as I can tell, the basic process the bus needs to support is this:

1. After the bus comes to a stop, exiting the bus is enabled, either automatically, or triggered manually by the driver. (you can’t have people jumping off the bus when it’s moving).
2. The ability to exit the bus is communicated to the passenger.
3. The passenger causes the rear door to open, and exits the bus.

Requirements 2 and especially 3 are the most interesting. Requirement 2 is filled on most buses by a green light over the door that illuminates. The location of the light is interesting, as in current designs it can be outside of the field of view of the passenger, sometimes causing confusion.

Requirement 3 is especially interesting, as this is where approaches vary dramatically between bus designs. The approaches I have seen (in chronological order) include:

1. A pressure sensitive top step. When a passenger steps on this the door opens.
2. A pressure sensitive bar on the door. When the passenger presses it the door opens.
3. A small motion sensitive region in front of the door. When the passenger waves their hand there the door opens.

All of these approaches have pros and cons.

1 pros: Works quite consistently. Most people can learn how to do it by observing somebody else doing it.
1 cons: Sometimes doesn’t work for small children, as they are too light. Doesn’t work on buses that don’t have steps. This second is serious, as all new buses are “low riders”. They don’t have steps in order to allow elderly and other mobility challenged people to get on and off easily.

2 pros: Also works consistently. Can be easy to learn for some.
2 cons: Can be difficult to learn for some. It is unclear that one has to _press_ the bar in order to open the door. The problem is that the shape of the bar does not possess the affordance of pushing. Also, the functioning of the bar is tied to the green light, but they are not in close proximity. It can be confusing as to why pressing the bar doesn’t open the door at first, but does a few seconds later.

3 pros: Doesn’t require physical contact. Can be important if there is a flu epidemic.
3 cons: The functioning is very unclear. First, the label on the door says “place your hand here to open”, but in fact that is not how it works. Motion is required. What happens is that a lot of people put their hand there before the green light goes on and press on the door. Unfortunately in that state their hand is motionless and the sensor doesn’t see them. So they end up pressing furiously on the door and nothing happens.

So, what are my conclusions? Option 1 worked quite well in the distant past. Unfortunately it simply does not work with the new low buses. And since old people and people with baby carriages should be able to ride buses, we have to forget about this. Option 3 is terrible. It is impossible to convey the functioning of the door sensor in a way that is immediately understandable.

Option 2 is the best, but needs to be tweaked. First, the bar that the passenger presses needs to be reshaped. Door designers learned long ago how to design a door in such a way that it is obviously either a “push” door or a “pull” door (although they sometimes sacrifice these lessons in the name of style). So, make the bar an obvious push bar, probably by making it wide and flat. Second, integrate the green light into the bar, possibly making it the shape of a hand. The green light will then indicate not only that you can open the door, but will also guide you in how to open the door.

Whenever I delve into development on the Facebook platform I run into new unexplored wtf head scratching territory. Today’s experience may take the cake. The last while I had a really weird persistent bug: Some AJAX server calls would get called by the Facebook server exactly three times, instead of one. Not two times, or four times, but three. Is Facebook a Monty Python Holy Grail fan? I figured not. I was probably calling something extra, somewhere, somehow.

Yesterday I put some effort into it, and found a single post on the developer discussion forums that explained the situation. It turns out, if your script doesn’t return anything Facebook will call the script three times. If it does return something Facebook will only call it once.

What?

You mean if my script causes something to happen on the server and doesn’t need to return anything to the app… a totally reasonable scenario, Facebook will make the decision of calling my script THREE TIMES?

This makes no sense, and as far as I can tell is not documented anywhere. Please Facebook, you are packed full of smart people. I know you can fix these things.

I don’t get why everyone raves so much about Windows 7. In my mind Vista wasn’t that bad, considering it’s a Microsoft product. And Windows 7 seems to be about as good. I had some problems with Vista that have been fixed (or mostly fixed) in 7, e.g. Bluetooth support, but new problems have cropped up. The first is that I can’t pin anything in the start menu. It lets me go through the motions, but nothing ever happens. Then, just now, it went and removed all my desktop icons. All of them are completely gone, except Adobe Reader. Yeah, screw you Microsoft. This was not my idea.

Some people have no idea what I do. Here’s a video of work that will be shown tomorrow at CHI 2010.