Not Quite a Blog

A friend, Dana Chisnell, is starting a kickstarter campaign, "Field Guides for Ensuring Voter Intent", to develop and produce a series of design and usability field guides for election officials. Dana, among many other things, is the co-author of the Handbook of Usability Testing we used at Berkeley when I was in IS214 (Needs and Usability Assessment).

The basic idea is that there are a lot of problems in human interaction with voting systems that can be easily avoided by a few smart design choices, low-overhead user testing of ballots, and other kinds of techniques and insights. Dana will spearhead the creation of a number of guides designed for election officials in mind that address common usability problems and techniques for avoiding them.

It looks like a neat project, and I encourage you to back it, if you deem it worthy. For funding levels higher than $30, you get a copy (signed) of each guide they produce... higher levels include hand-made art of a mascot for the project, "Chad Butterfly"... the uppermost levels of funding allow backers to pick an election official of their choosing to receive a consultation with Dana about usability aspects of their election process.

This is a neat example of both civic design and indie usability funding! Bravo! I hope Dana meets and exceeds her funding level!

Make no mistake about it, Louis C.K. is one of the funniest humans alive.

The web has been abuzz this last week about an experiment Louis C.K. performed where he released his latest comedy special at a low price and without DRM. Tuesday, he released a statement where he described the results of the experiment: in short, he broken even after only twelve hours.

His statement contains a lot of compelling rhetoric about cutting out the middle-man and delivering a more fair and flexible experience to his fans. However, when I went to go buy the video, I was surprised to find the terms and conditions are very different from the rhetoric of his statement.

Louis says a lot of things in his statement, but a few things stood out to me. He sought to (1) lower prices, (2) make purchase and viewing easy, (3) eliminate restrictions on viewing, (4) reduce incentives to "steal" content, (5) not use purchasers' personal information as an asset, (6) promote collective ownership of the product, and (7) eliminate contacts and marketing from the producer.

Let's look briefly at each of these in turn...

Full story »

NYC's ABC7 did an awesome spot on Scott's Pizza Tours, my favorite local activity. It's a great piece that describes well what you learn and experience on one of Scott's tours. I'm briefly in it and say a silly quote, trying to emphasize the sheer variety of different kinds of pizza that drives us pizza enthusiasts.

Enjoy!

(cross-posted at the Privacy Research Blog.)

An intriguing story flew past my Twitter stream, that begins:

"MINNEAPOLIS (WCCO) -- Detailed medical information discovered on the back of a first-grader?s school drawing sent Minneapolis school officials scrambling.

Jennifer Kane was tidying her dining room when she found the drawing by her daughter, Keely, who goes to Hale Elementary School. On the back of the paper was the name, birth date and detailed medical information for a 24-year-old St. Paul woman named Paula White." --("Recycled Medical Records Used As Scrap Paper At School")

Long story, short: Ms. White's records that she voluntarily gave to a law firm representing her after a car accident were donated by a paralegal to Ms. Kane's daughter's elementary school. These records, and those of presumably many others, were found by school officials after being used as scrap paper and have since been secured, probably waiting disposal (or, cynically, placed in escrow until the new team of lawyers Ms. White might hire to sue her old lawyers get a chance to look at them!).

Ms. White expresses concerns that we see often in cases of privacy breaches, especially medical breaches: "It?s got my account number, my birth date, my job ... I?m outraged. I am embarrassed. I don?t want anyone to know my personal information.?

What recourse does she have? Well, because the law firm is not a "covered entity" under the federal law and accompanying regulations known as the Health Insurance Portability and Accountability Act (HIPAA), the responsible enforcement agency, the department of Health and Human Services, can't seek corrective action. In fact, you may be surprised how little HIPAA and HHS can do in situations like these. Our friends at the World Privacy Forum keep a very useful FAQ about HIPAA and also point out how medical identity theft, where people use medical information about others to obtain services or make fraudulent claims, is on the rise and an increasing concern for patients.

What can you do? Be vigilant, as always. Make sure you monitor and understand your health insurance claims information and that you let your health care providers know if you suspect funny business.

Since I've started research involving personal health records (PHRs) and privacy, I've gotten interested in health "gadgets", like the fitbit. The fitbit is a small device--with an accelerometer and altimeter in it--that measures how active you are throughout the day.

I bought one and immediately noticed it must have been inaccurately calculating the distance I walk/run. For example, I tweeted the other day: "@fitbit says my morning run was 6.74 mi., when the Gmaps pedometer www.gmap-pedometer.com says it was 4.44 mi. 2.3 mi. difference!"

What likely had happened is that I messed up: I had initially entered default-like values for my stride length in the fitbit website settings and then removed those settings. Fitbit says that if you have no stride length entered, they calculate your stride length based on your height and gender. Being over six feet tall means that my stride is a bit long, but something must have been wrong: fitbit thought my stride was much much longer than it is and attributed to me much more distance and speed than I actually put out!

I set about to recalibrate my stride length and running stride length and used the following procedure, to good effect. These instructions are easier to follow than the detailed version from the fitbit website, and don't suffer the flaw of the FAQ version that mixes up numerators and denominators:

1. During a normal walk and then later on a normal run, log a fitbit activity of a well-marked distance (hold down the button for a bit until it starts counting like a stopwatch, then do the same to end the activity). The distance should be over 1/4 of a mile and should be marked by structures easily seen on Google Maps' satellite view (you'll see why in a bit). I'd suggest going for a 1 mile walk and recording an activity on the way out and the way back for about 1/4 of a mile in between. You should walk and run at a normal pace (not fast, not slow).

2. Upload the activity data to the fitbit website and note the amount of steps taken during each activity. For me, I had one walking activity recorded between a bridge and a conspicuous house and two running activities (at the beginning of a 5 mi. run and then near the end). It came out to walking 619 steps and two running readings at the same distance of 524 and 536 steps (average of 530 steps).

3. Point your browser to www.gmap-pedometer.com to calculate the distance between the beginning and ending of your calibration distance:

   1. Enter in the zip code of where your run was and zoom in as far as you can to where you think the structure that marks the beginning of the calibration distance would be.

   2. Turn on the "hybrid" view mode so that you can see the street and the satellite image.

   3. Hit "start recording" and check that "Draw route:" is set to manually (straight lines).

   4. Double-click where your feet were when you started your activity to start the path recording.

   5. Turn off "hybrid" view for the "map" view and click out the path of your distance.

   6. When you get to the end, turn "hybrid" mode on again and make sure your path ends as close as possible to the structure marking the end of the calibration distance.

   7. This distance, in my case 0.3192 mi., is the calibration distance we will use to calculate our stride lengths. You might want to save that gmaps-pedometer route for future reference.

4. Now we do some basic math:

   1. Round the distance: While gmaps-pedometer can zoom pretty close, I'd suspect that the path distance is off by +/- 5 ft. at both ends simply due to the limited resolution. So, round your distance to three decimal places (about 5 ft.). In my case this makes it 0.319 mi.

   2. Convert distance to feet: Convert your calibration distance to feet, of which there are 5280 ft. in a mi. In my case: 0.319 mi. * 5280 ft./mi. = 1,684.32 ft. Which is about 1,680 ft. when factoring in significant figures.

   3. Calculate stride lengths: Divide this figure by the number of steps:

      • 1,680 ft. / 619 = 2.71 ft. = 2 ft., 8.52 in. (walking)
      • 1,680 ft. / 530 = 3.17 ft. = 3 ft., 2.04 in. (running).

5. These last two figures are what you should enter into the fitbit profile settings for stride length and running stride length.