Much has been written about the impending death of newspapers. The wealth of information on the Internet is of course threatening the existence of newspapers, where what you are reading is filtered by editors that you may not agree with, can lag actual events by up to 24 hours (!), and costs money.

I actually like reading newspapers. Once or twice a week I will relax at lunch and go through either the National Post or Globe and Mail. It gives me an opportunity to actually cogitate, as opposed to the web experience which, to me, is more frantic and random.

The point of this post, though, is how newspapers are trying to build online presences. Both of the mentioned newspapers have online versions with extra functionality, such as the ability for readers to write comments. This is very nice. Today I was looking at the Post’s website, and ran across a particularly ridiculous article that deserved to be commented on. I felt compelled to do so.

The first difficulty presented was the fact that I had to sign up to comment. This is an unfortunate necessity, as no accepted standard for universal online ID management has emerged, and they need a mechanism for filtering out the vi@gra comments. So, I filled in my info and submitted it. I then tried to comment, but discovered that I had to “upgrade” my account to be able to participate in discussions. This is a bit strange. I’m not quite sure what the point of an un-upgraded account is, as it doesn’t seem capable of doing anything. Anyway, I pressed the “upgrade” button, and discovered that I had to click on a link from an email that they sent me, to prove that my email address is correct. I clicked on the link, and the resulting website said “E-Mail Address Not Found.” So, I must confirm my email address to comment, but I can’t confirm my email address.

I suspect the problem is that the email address I gave them is a Gmail address with a “+” in it. Using a + in Gmail is a technique for filtering incoming email from places you’ve signed up. My suspicion is that the Post’s email address parser has trouble with the “+” character.

Anyway, the outcome was that I wasted 10 minutes, still can’t write comments on the article, have another email address floating around out there in their system, and have a little less respect for newspapers’ abilities to move beyond their current business model.

For the last several months the driver-side power window in my car has been sluggish going up and down. So, yesterday I spent 4 hours replacing two plastic guide bushings hidden in the innards of the door. The bushings are really simple tiny things, and cost about 50c each, but in order to put them where they are supposed to be one has to remove several layers of door material. After struggling through the process I realized that there is a whole unexplored area of usability research just waiting to be investigated: mechanic car interaction.

Researchers have spent a huge amount of time studying how people actually drive cars. We have a very deep understanding of how to design cars from this perspective. But I have never heard of any research centered around the maintenance of cars. I think this is a very interesting problem. On one hand I want car maintenance to be so trivial that fixing any problem is like building a simple LEGO model. Everything goes where it seems it should, and clicks nicely into place. However there are a bunch of competing requirements. The car must remain affordable to build. It must be light enough to have reasonable gas mileage. It must be safe. So, can we design a car that can be maintained by anybody but maintains these properties?

I think the answer is yes. Taking my door as an example, the entire thing contains maybe 50 parts. That is not a high degree of complexity. Once the thing is apart it is not terribly difficult to figure out what does what. The difficulty lies mostly in actually getting it apart. The fasteners used are generally not friendly. To get the first body panel off one must tug at it, fearing the whole thing is going to tear apart, until the fastener finally lets go and pops out of its hole. Then there is a layer of vapor barrier glued on with the grossest sticky black stuff. Then there is the air bag that has to be treated gingerly, lest it blow up in your face. There are bolts that fit into little tabs that tend to break. Finally, during reassembly, there is waterproofing trim that is a freaking nightmare until one figures out that it needs to removed and fit on the floor panel, rather than trying to stick the door panel onto it while it sits on the door frame. All very confusing.

In short, I think a huge difference in maintainability could be made by improving fastener technology, and the way in which parts fit together. Taking the door panels off should be a pleasure. I should be able to lift a lever and have the things slide right off. The camera makers have figured this out. Sometimes I take the lens off my SLR and then put it right back on, just because it feels so nice. Why can’t cars be the same?

Looking beyond door maintenance to more standard car stuff, why do I have to stick my fingers into a dark and dirty hidden place and pull on a sharpish lever in order to raise the hood? I read somewhere that some non-insignificant fraction of drivers never lift their hood. I suspect they would be more likely to if lifting it was much easier. Also, when adding oil why do I have to open this bottle and then pour it into this smallish hole, potentially spilling it all over either the engine or myself? Oil is very unpleasant stuff. There should be a method for adding oil where the oil itself is never arcing through empty air, potentially making a mess.

A while ago I read about a team of average women who were recruited by a car maker to design a “woman’s car,” whatever that is. One of their answers was to make the hood impossible to open, except by a licensed mechanic. I think this is an example of the mistake of letting users design something. They designed what they thought they wanted, not what they actually wanted. And since they didn’t like opening the hood, their answer was to never do it. What they actually want, and what I want, and what I think everybody wants, is a car that lets me change pretty much any mechanical part (maybe stopping short of the powertrain itself), without getting dirty, pinching my fingers, or worrying about breaking anything. Maybe the women from the story won’t go so far as to do the maintenance I did this weekend, but maybe they will be able to open the hood and put some oil in. The point is that the difficulty level of every aspect of car maintenance, from adding oil to changing a bushing, can be made simpler.

As for my adventure with the window, it now opens and closes like it did when it was new. And I only stripped one bolt, and ended up with one mystery washer.

I always derived some perverse kind of pleasure from doing my taxes. For years I insisted on doing them by hand, and always felt some satisfaction when I made everything add up nicely. This was until, two years in a row, I made an error early on in the process, which forced me to redo large swaths of my work. The joy was gone, and I decided to no longer do my taxes by hand.

Now, like many people, I have adopted online methods. I personally use www.ufile.ca. This is a surprisingly well-designed, user-friendly system. One can punch in the numbers from their various tax slips in any sort of order, and it will instantly provide you with an up-to-date result at any time. But, there is a hidden dark side to the ufile tax preparation approach.

Unfortunately, ufile might just be TOO easy to use. Recently in preparing my taxes I ran through my slips in no time, clicked through to the results page, and was happy to see that I had a modest refund owed me. Then I saw something odd. First, I hadn’t used any of my available education amount for the year. Second, the result page didn’t show any capital gain, whereas I expected to have one. After poking around I found my error. I had keyed in some digits wrong for one of my capital gain entries, and instead of claiming a $1.06 gain for the trade, I had claimed a $960,000 loss. This is a substantial error that explained the observed oddities. What scares me is that it was so easy to make this error, and simultaneously difficult to spot it once it had been made. If I had been doing my taxes by hand I would have definitely caught this, but since ufile is so easy, these kinds of errors are simple to make.

I suspect that, compared to the old days of paper returns, there are a large number of people who are making the kinds of mistakes I describe. It is too easy for a mis-key to result in 1 or 2 extra digits somewhere. But I think the problem can be addressed. What the ufile system needs is some kind of intelligent “Clippy” equivalent that will do a sanity check on your inputs. On my return, for example, it might look at the numbers in aggregate and ask me “Hi Garth! Say, you don’t seem like the kind of guy who would lose $960,000 in a money market fund. Are you sure this is right?” Or, in the case of another mistake that I caught, it would say “Hi Garth. I find it kind of strange that these two different trades have exactly the same numbers entered for ‘cost’ and ‘proceeds’. Are you sure you didn’t enter the wrong numbers in one of them?”

So, in conclusion, ufile and web tax preparation services in general are great. Their main problem is that they are TOO great, making simple errors easier to make and overlook. There needs to be a mechanism to address these kinds of errors.

I often give Microsoft a hard time, and justifiably so, I think. But that doesn’t mean that everything is roses in Apple land. While Apple’s products are generally of the superior sort: my G5 tower is built like a tank and functioning perfectly 5+ years on, there are flaws that occasionally bite me in the ass.

The first problem is Apple’s obsession with being ultra cutting edge. This started with the elimination of the floppy drive in the iMac, which was controversial but turned out to be prescient. That didn’t cause me problems, but other changes have. My new Macbook Pro, for example, only has a Firewire 800 port. This causes problems when my audio interface only has a Firewire 400 port, and the book store is out of adapters. I end up instead recording on a more ancient Mac and then having to copy the audio over to my new one. Tedious.

Then there are the video outputs. I had just started to get used to DVI, and then Apple came out with these mini-DVI things, or whatever they are called. I’m not even sure. There are even a couple other standards that they supported that I skipped over because they seem to change every year. A couple of weeks ago in a meeting a professor was unable to project because her new Macbook adapter was DVI-D and didn’t have analog out that could feed the projector.

From a developer’s perspective there is Apple’s bizarre choice of going with Objective C as the language of choice. At the time of this decision I don’t think I had ever actually met anybody who knew Objective C. I think the choice had something to do with their purchase of Be and rolling BeOS into OS X, but still, weird.

I still think Apple makes some of the best hardware and software out there. It is just a pain sometimes when you are hauled kicking and screaming into the future, when you just want to sit back and enjoy your Firewire 400 device for a few more years.

Today, in a rather desperate attempt at improving the performance of my demo by running it on a different machine than usual, I went about installing MS Visual Studio on a machine that was otherwise Visual Studioless. Microsoft is nice enough to provide a free 90 day trial version, so I downloaded that and went about installing it. Except I couldn’t, because they distribute the installer as an .iso file. So what’s the problem? An .iso file is a disk image, which Windows is able to burn to a CD. That’s fine enough, except I didn’t have a CD handy, and even if I did I wouldn’t want to waste it on burning the .iso. After all, why do I have to move the bits from my computer to a CD just in order to get them back onto my computer? They are there to begin with!

And then I discover, shockingly, that Windows XP doesn’t know how to mount an .iso. I would think this was basic functionality, and it is functionality provided by a number of non-free third party software suites, but Windows doesn’t know about it. So, Microsoft is distributing a free trial version of its development software, trying to attract developers to its platform, and is distributing that software in a format that is nothing short of a gigantic pain in the butt.

Eventually I discovered that there is an unofficial, unsupported .iso mounter available from Microsoft. I imagine somebody internally sat there trying to install Visual Studio (or something similar) and thought “This is bullshit. Why the hell can’t Windows mount an .iso?” So they went and wrote a utility to do it.

And of course, on a totally different topic, Apple’s developer tools are entirely free.

Oh, and as I was writing this my Windows machine went to sleep, in the middle of importing a video from my video camera. Thanks Windows! I didn’t actually want that anyway! And the importing application crashed in the process (the second time in the last 1/2 hour). As Christian Bale would say: “you’re a fucking amateur, you know that?”

I usually use a Mac at school, but I often do development on Windows, so I ordered a Dell XPS laptop. I received it today, after almost a month of waiting. Here I will catalog some of my observations on the Windows/Dell user experience. This list will hopefully expand as I notice new things.

1. The packaging is pretty cheap. The box is twice as big as that for my MacBook Pro, and things seem just randomly tossed inside.
2. Windows welcomes you in a most unwelcoming way. On the Mac it plays you a friendly tone and shows you a snazzy video. Not exactly useful, but the gesture is appreciated. On the Windows machine pretty much the first thing you see is a dialog box saying “click here if you agree with this end-user license agreement.” Twice. That’s like walking into an ice cream shop and getting kicked in the groin.
3. The IE title bar says “Internet Explorer provided by Dell.” Listen Dell people: I know I am using a Dell. You don’t need to put your company name in the IE title bar. It reeks of desperation.
4. External monitor support is clumsy.
5. The build quality is actually pretty good, although the track pad is small and its surface texture is lacking.
6. Please don’t put stickers on my laptop.
7. My first time logging in there was a flurry of various security programs complaining about other security programs.
8. The default items in the Windows menu are: Windows Media Center, Windows Ultimate Extras, Windows Media Player, Windows Photo Gallery, Windows Live Messenger Download, Windows DVD Maker, Windows Meeting Space, and Windows Mobility Manager. This is branding run amok, and somebody should be ashamed for putting the word “Windows” 8 times for absolutely no reason in the most important menu in the OS.
9. The Google widgets thing crashed the very first time it ran, then the second time it ran, then the third time it ran. It will be run no more.
10. I suppose you think you’re cool by making the multimedia controls entirely touch-based (i.e. no tactile feedback) but in fact you are just being dumb. Kind of like the kids in highschool who wore their pants backwards to be like Kriss Kross.

I have a Chapters gift card for $25. On the back is the presumably unique bar-code for the card, and the numeric equivalent for the bar-code written underneath. The number is 31 digits long. That makes possible roughly a nonillion (thanks wikipedia!) unique gift cards. If Chapters sells a million gift cards a year, they won’t run out of numbers until roughly septillion years from now. This puts my credit card, with its mere 16 digits, to shame.

So, what’s the big deal? Who cares? They will just scan the card through the cash register anyway. You don’t care about the numbers on the back. Well, it turns out that you do if you want to buy The Black Swan and Use of Weapons using the Chapters website. To use your gift card on the web you have to type in the 31 digit code by hand, and then when it turns out to be incorrect you have to type it in again, and then a 3rd time. And then eventually, after typing in 31*n (n depending on your tolerance for repetitive tasks) numbers, you give up and swear at Chapters.

So, what’s the lesson? First: 31 digits is a few too many. Second: there should be a mechanism for using gift cards on the web that doesn’t involve typing in numbers at all.

I’ve never been a big fan of the whole “crowdsourcing” theory as it relates to the wisdom of crowds. I am rather a believer in the power of the individual, a single person’s ability to achieve great things. This may reflect my own hope that some day I will emerge as being one of these great people (I’m still waiting), but I think there is also a fair amount of evidence pointing towards the over-hyping of crowdsourcing as an approach towards accomplishing great things that are not achievable by small talented groups.

My favourite example comes from a talk at a CHI conference (2007 or 2008), where some guy was giving a talk on the subject. His best example of successful crowdsourcing involved three composers who collaborated in writing an opera. First of all, three is not a crowd. Second, these three were not members of “the masses,” but were rather expert composers. This example is a big fail in my book.

This leads me on a bit of a tangent to thoughts on online communities, where discussion forums serve as collections of wisdom created in real time by crowds of collaborators. The two examples I will use are www.digg.com and www.fark.com. Both of these communities are centered around the discussion of news items, but I have observed that the “collected wisdom” produced by the two communities differ vastly in quality. In my opinion the discussion on digg tends to be simplistic, poorly thought out, and often of little value. In contrast, the discussion on fark is often sarcastic and silly, but it is at least clever, and this sarcastic content is intermixed with a good deal of intelligent commentary (at least in comparison to digg). Why are digg and fark so different?

I believe the quality of discussion on digg and fark differ so dramatically because of the reward structure in place on the two sites. digg has a complex system for allowing users to vote for comments and stories they like. Higher scored (more popular) comments receive more exposure, and lower scored comments are hidden. This is meant as a mechanism for encouraging quality content. On the other hand fark is egalitarian, in the sense that a comment written by any one user is given exactly the same weight as a comment written by any other user. I believe that the digg approach works exactly opposite to its intended purpose, and the fark approach that would seem to invite abuse instead encourages quality. The problem with the digg approach is that many users are driven primarily by the need for the approval implied by high scored comments. Instead of writing what they think to be interesting and of value, they write what they think will be voted up. These two things are not the same. On the other hand there is no reward structure on fark. There is only the discussion. There is no secondary market of “approval points” to worry about, and I believe this keeps the discussion honest and interesting.

Is there a take-home lesson to this somewhat meandering post? First, I don’t think a crowdsourcing approach will ever be effective at producing breakthroughs, but it can be effective at achieving less grand goals. Second, when constructing a crowdsource-based community one has to be very careful designing the reward structure, because the results you achieve will be optimized for that reward structure.

When the iPhone was being rolled out, what really made me perk up was the announcement that there would be an app store where developers from around the world could sell their apps. My reaction was excitement. My thought was that the app store could really differentiate the iPhone from other phones, and could foster the kind of creativity and community that elevates a technology from mere tool to life-changer.

It seems now that the app store is living up to my expectations, but there are clouds on the horizon. I just read today that the current best-selling app is “iFart Mobile.” If you are unfamiliar with this app, it allows you to reproduce a number of distinct fart noises (e.g. “Bubbler” and “Kazoo”) for the delight of friends and family. This has me worried.

My concern is that Apple will make the same mistakes as Facebook. We all know what happened with Facebook apps. The quality Facebook apps were overwhelmed by the crappy garbage that was designed specifically and exclusively to be viral. The Facebook app community is suffering as a result, and I am concerned that the iPhone app community will likewise suffer.

I don’t have a solution to this. All I know is that Apple has to somehow cultivate quality iPhone apps, while discouraging the crap. Allowing the community total power to do so is not effective, as the clever app creators will be able to manipulate the community. Apple must take an active role, while simultaneously respecting community input. Striking a balance between direct intervention and a hands-off laissez-fair approach is a tough thing to do, but is critical, as I think the vitality of the app store will largely determine the longterm fate of the iPhone.