A Broken iPhone App in the Wild
Around November last year, I embarked on a project that I may someday regret. I started programming, to build an iPhone app. Then, it morphed from dabbling to an obsession. I originally allotted no more than 10 hours a week to the project, then, for two months, I pour nearly every minute of free time into it.
The idea for the app started from a conversation with my father, who did some research in speech technology for IBM. Text to speech, or speech synthesis, is an old concept. Computer could read text out loud for decades, but most of us remember the robotic voices associated with bad science fiction. Over the years, speech synthesis has steadily improved, but it found little practical use except specialized applications like automated response by phone for customer service and devices like GPS. Is there a way for the average person to make use of it?
With the rise of mobile computing, text-to-speech takes on new importance, because smart phones have small screens but powerful computers inside. Technically, it’s possible to have very good text-to-speech on an smart phone. But what content would an average person want to listen to? As I looked at text-to-speech apps for the iPhone, what I found missing is the acquisition of content. All the text-to-speech apps are a bit idiosyncratic: “Have your phone say something silly as a prank to your friends…”
If many people want to listen to the same piece of text, like news or a popular book, you can always find audio versions read by real human beings. Human voice will always be superior to computers generated voice. So the kind of content fit for text-to-speech is the kind of content that nobody would find it economical to record a real human being reading it. The obscure stuff, like professional journals, blogs, papers, etc.
I’ve been using Instapaper on my iPhone and iPad for a long time. When I see a good but long article in a Web browser, I’d mark it to “Read Later”. Typically, before I get on an airplane, I’d download these articles on my iPad and iPhone. Then, during the flight, I would be able to catch up on my reading. Additionally, Instapaper stripes out the advertisement and other extraneous stuff on the webpage, leaving just the juicy content. The idea of combining Instapaper and text-to-speech came to me.
But what kind of people would want these articles read to them, rather than reading them on a screen? Commuters, of course. They spend a lot of time in a car, like I spent a lot of time on airplanes. Unlike an airplane, a driver can’t read on screen (except for a few people I used to work with — always caused great terror when I sat in the passenger seat). So I have the target segment: commuters who want to listen to articles not meant for mass consumption.
This idea had been stewing for a while until the dark arctic winter during which I needed to do something while it was dark outside. I had not done any serious programming for 18 years, so I was rusty. Very rusty. Technology had changed in a big way. It was like I knew how to ride a bicycle, but I had to figure out how to ride a motorcycle.
When the app was finished around the end of last year, it not only could acquire content from Intapaper, but also another popular bookmarking service, Read It Later, and it could acquire content from PDF files, text files, and cut-and-paste. It had a playlist feature, so a driver can set up a list of articles to read during a commute. The content is stored locally on the iPhone and iPad, you can listen even if you lose Internet connection.
Two days ago, it got on the Apple App Store. It’s been out for 2 days, during which 100 people bought it, quite surprising considering I did no promotion. Yesterday, I got an email from a woman in Sacramento, who said that she couldn’t get Read It Later to work with the Reader. I asked her to checked a bunch of things without avail. I was stumped. Today, I started to get a terrible, nagging feeling that something may be very wrong. So I pretended to be a customer and bought to app myself from the App Store. I downloaded it, and lo and behold, it wouldn’t work with Read It Later. I wouldn’t work with Instapaper, either. It was a moment of sheer panic. I put this piece of software out there and it doesn’t work! I furiously went through the code to find out what could go wrong. After a hour, I found it. One stupid line of code. When I built the app for testing, this line of code got executed. But when I built the app for the app store, this line of code got ignored. I changed the description of the app to tell people that this feature doesn’t work, and a fix is coming next week.
People ask, so this means you’ll stay on Rødøy? No. This is a hobby, not a career. I don’t see how this app can ever generate enough income to support my family. But it’s fun for now, when my wife is supporting the family. Thanks wife.