Learning Chinese is not for the faint of heart. Not only does the non-native Mandarin speaker have to master the language’s infamous tones, he or she must memorize thousands of (practically speaking) non-phonetic characters, get acquainted with a range of accents, and grapple with a deceptively simple grammar system.
At the same time, even the most gifted linguist will admit that one of the biggest challenges posed by Mandarin isn’t the mechanics of the actual language, but the grunt work required to learn it well. Looking up characters in a paper-bound Chinese dictionary is a multi-step process that can take tens of minutes if you’re not careful. Also, relying on a single Chinese-English dictionary for reference is a surefire way to commit language suicide. For such a long-lasting, quickly-evolving language, you’ll need at least three dictionaries handy in order to get a rough idea of what a specific character, word, or phrase means – and even then you’ll usually have to apply some brainpower to figure out how it’s used properly.
Enter Pleco – the best Chinese dictionary app on the planet. To some of our readers, a dictionary app might not seem like the most exciting of subjects, but those who know and use Pleco understand how crucial it is to one’s language learning regimen. It’s one of those rare brand names (if you can call it a brand) that will elicit sheer glee from its users upon the very mention of its name. A Swiss Army knife app featuring 25 dictionaries, almost anyone that’s used it can recall a moment when Pleco “saved their life.”
While the app has won legions of fans, few are aware just how revolutionary it was and continues to be. Pleco was first launched as an app for Palm in 2001 – before the big boom in Chinese language learning and the world’s mass adoption of mobile handsets. It pioneered the notion of a Chinese dictionary as a powerful, always-on tool for a wide range of learners, and was the first cross-platform Chinese dictionary to merge handwriting input with character searches across multiple dictionaries. Want to know what 熊貓 means but don’t know how to pronounce the characters? Just trace them in the input field and you’ll find the word next to “panda,” its definition, alongside “xiongmao,” its romanized phonetic pronunciation.
Now the app features optical character recognition (“hover-to-translate”), mixed character-pinyin search (trust us, that’s a big deal), voice input, flashcards, and many other bells and whistles that make learning Chinese that much easier for hardcore students.
When you consider that for centuries, the only way to look up the word for “panda” was to count the number of strokes for the radical component of 熊, consult a series of charts, and then hope that the suggested definition remotely made sense, the convenience of Pleco marks a major turning point in the history of Chinese language learning.
Moreover, more than ten years after it first appeared on Palm, Pleco remains a mostly one-man operation. For 32-year-old Mike Love, a programmer based in New York, Pleco is a full-time hobby that doubles as a business. While many of the apps on our smartphones were created by Silicon Valley hustlers with pipe-dream ambitions and half-baked business plans, Love has added tremendous value to language learners around the world simply by building a better dictionary. Think of him as the pastor overseeing the long-awaited wedding between the Chinese language and mobile electronic devices.
In an effort to learn more about his role as the unsung hero of Chinese Learning 2.0, Tech in Asia sat down with Love to talk about the genesis of Pleco, the evolution of mobile technology, and the perks of solo entrepreneurship.
Let’s start from the beginning. What are the origins of Pleco?
In my senior year of high school I was studying abroad in China in 1999. I saw these neat little portable electronic dictionaries that every Chinese person was using to learn English. I wanted my own version of one of those dictionaries, but for Mandarin. I didn’t have a factory in Shenzhen to churn out my own portable Chinese dictionary, so I needed an already-available device to put it on. At the time, that device was the Palm.
My parents gave me a Palm IIIx for Christmas in 1999. I put something together using some off-the-shelf Palm apps – a handwriting app, a Chinese enabler app (since Palm wasn’t Chinese enabled), and CEDICT – a predecessor of CC-CEDICT – which was a simple Chinese dictionary that had about 20,000 entries. I built a prototype using those, and then some of my friends started buying Palm Pilots because they wanted to use it.
Around March 2000 I wrote to Oxford University Press, which had a really awesome Chinese dictionary at the time, and asked, ‘Is there any way I can get an electronic copy of your dictionary?” Back then, I wasn’t planning on building a business out of Pleco, I just thought that if I could put Oxford’s Chinese dictionary on my friends’ Palms, it would be a nice thing for them. They said “Sure, we’d love to license you our dictionary!” I had no idea it would be that easy. So we negotiated some terms and then we had an exclusive copyright license to the Oxford Chinese dictionary for Palm Pilot. Pleco was founded two weeks after my 18th birthday in May of that year.
After you put the Oxford dictionary on Palm, you went off to Harvard to get your Bachelor’s in Computer Science. During your time there, did Palm remain an ongoing project for you? Or did it shift to the backburner?
I kept working on it. I had to, because it was successful. We officially launched in October of 2001, and it was doing okay, making me enough money to keep my social life in good standing. Then That’s Beijing did a piece on us in April of 2002, and our sales doubled that month and basically stayed at that level moving forward. I started making more serious amounts of money and got to thinking, ‘Gee, this could actually be my job after I’m done here.’ My dad was a high school principal, and my mother was a university admissions officer, so I was never going to be allowed to drop out of school, however tempting that might have been! But I was doing well enough that I saw Pleco as an option I could pursue if I didn’t see anything that I liked more. So I kept it going.
I graduated, and did not in fact find a job that I liked better than Pleco. I completed a summer internship at Microsoft, but I didn’t really enjoy myself much there. So I thought, ‘Okay, I’ll just work for myself.’
Why didn’t you stick around at Microsoft or work at another large firm? You would probably make a great candidate for a job at Palm!
The stuff that the high-level gurus at those places do is math-y theoretical computer science that I don’t really find much fun, quite frankly. I did reasonably well at it in school, but I never really enjoyed it. I don’t really like spending a lot of time thinking about Big O notation and so forth. And at the lower-rung levels, it wasn’t very interesting either. You could work on a tiny sliver of Windows, and maybe if you try really hard in 20 years you might get to be some sort of development lead. And even then you’d still be managing a little piece of someone else’s ecosystem. Even if you’re working at a big company with career security and free soda, you are really just a tiny little piece of something. I wanted to run the whole thing, even if it was just a small thing.
Pleco is probably one of the few mobile apps in existence that’s continued to thrive since the Palm era. Can you walk us through the product’s evolution as smartphones have grown more commonplace and more powerful?
Our first killer feature for v1.0 was handwriting input, which we launched just as touchscreens were coming out. Most people didn’t even know that touchscreens were a thing at that point. I had seen a few on some of the high-end electronic dictionaries that were popular in Japan and Korea, and I knew it was something we had to have. I saw it as the first step in making something significantly better than a paper dictionary.
Building the handwriting input was a case of good fortune. Motorola had a really good Chinese handwriting engine, so I spoke to them about licensing it. Most copyright licenses involve some sort of royalty advance, or at least a commitment to buy X number of units minimum over the term of the contract. Motorola initially offered us a license with a much larger commitment than we could afford, so we countered with double their per-unit price but no up-front commitment. Surprisingly, they agreed to that. I don’t think anyone was expecting anything out of Palm Pilots at that time, so they were willing to take whatever money they could get.
Back then, it seemed to me that Motorola and Oxford were very generous about licensing terms – more willing to accommodate a bootstrapped startup of very limited financial means than copyright licensors tend to be with mobile app developers now. Both Motorola’s and Oxford’s licensing regimens back then were mainly tailored towards OEMs. Both were working with companies that made the sorts of standalone electronic dictionaries Pleco was inspired by. I don’t think they expected downloadable third-party software to make much money. But for that same reason, they also didn’t see it as a threat to their OEM customers. They could license to us and make a little extra money without giving up any of the money they were getting from OEMs.
How did the spread of the mobile web affect Pleco as a business?
The mobile web wasn’t even a consideration in the pre-iPhone days – people might have had access to it, but it was always too slow or too expensive to make OTA distribution feasible. Pre-iPhone devices also generally didn’t ship with Chinese fonts so web-based competitors weren’t really a concern either.
Post-iPhone, it’s certainly been a boon for distribution – an awful lot of our business comes from word-of-mouth, so the fact that somebody can hear about Pleco, pull up the App Store and download it instantly is tremendously useful. The promise of that was a big part of what motivated us to adopt the free-with-in-app-purchases business model so early – Apple initially restricted in-app-purchases to paid apps, and our plan at that point had been to offer a ‘basic’ version of Pleco for US$10 or US$20 and sell various additional dictionaries as in-app purchases. But in mid-October of 2009 Apple opened in-app purchases to free apps, and we quickly restructured everything around that so that when our app launched in mid-December, we were able to make the initial download free.
And of course, in-app purchases themselves have also been a huge boost – for our Palm and Windows Mobile apps you not only had to visit a website to buy our software, but once you’d bought it, you had to copy a “keyfile” from your computer to your phone in order to activate your purchase. So the first part of that cost us a lot of sales and the second half of that left a lot of our users frustrated spending hours trying to get their newly purchased software working.
It seems like if Pleco were in the hands of anyone else but you, it would have died. Here’s a guy in his early twenties who launched a DIY Chinese dictionary app for mobile. One would think that demand for a talented mobile developer, even in the early 00’s, would be high, while demand for a mobile Chinese dictionary in that era would be relatively low.
Well, you’d be surprised. I try not to discuss numbers in too much detail, but I think the profits were in the six-figure range by 2005-2006. So I was certainly making something competitive with a senior developer salary just on Palm and Windows Mobile. It’s not like I could have gotten a job that paid me more.
Pleco was niche, yes, but we had something that was unique enough that people were basically going out and buying Palm devices just for Pleco. The second-most popular page on our website, after our homepage, was our “What Palm do I buy to run Pleco?” page. So even though we had no hardware, people were buying hardware just for our app.
One might also assume that the most obvious path to growth for Pleco would be to venture out into other languages. Have you ever at any point considered launching a like-minded app for Japanese? Or Spanish?
I once tried to to develop a phrasebook product for a few different languages, which flopped spectacularly for reasons that should be obvious. A phrasebook is about the size of a Palm Pilot, it costs about US$8, you can get it dirty, you can throw it away after a trip… Why would you want a marginally easier to browse Palm version of this? You still don’t see a lot of successful phrasebook apps, even on the iPhone. Phrasebooks are great – they’re durable, and it’s one case where paper is probably superior. So that was my one attempt to branch out into other languages.
But even then, I wouldn’t say that the failure of that attempt led me to stick to Chinese. The reality is that I just found Chinese much more interesting. I never run out of stuff to do with Chinese, it’s such a rich and incredibly complicated language. So I never really saw the point of expanding horizontally into different languages.
What do you mean when you say you “never run out of stuff to do with Chinese?” What are some of the qualities unique to the Chinese language that keep you busy?
Well, character reading for one thing. You’ve got a language where really, really experienced Chinese learners still can’t read all of it. So characters present all of these problems – how do you teach them correctly? What are some better ways of indexing them? Moreover, the fact that Chinese learners fundamentally can’t read the characters that they need to look up in a dictionary means that there are problems you have to solve regarding input methods – how do you search a character that you don’t know? With Chinese, unlike other languages, there’s an ongoing need to work with the actual text.
In Spanish, if you don’t know a word that’s in a book you’re reading, you can type that word into a dictionary very efficiently. With Chinese, you’ll have to go through some hoops to look up that word. It’s the only language where you’ll never be able to read all of it – there will always be a gap between what you can read and your knowledge of vocabulary.
A few weeks ago on Twitter you said that you weren’t interested in pouring resources into Pleco’s OCR technology, which has long been hot topic in the Chinese language learning community. If not OCR, what elements of mobile app-enabled language learning excite you?
I’m interested in the gamification of everything. I love Foursquare, and I think there’s a lot of potential for language learning to provide a quantified, reward-based way to keep you motivated. I like the idea of integrating Chinese language learning into your daily life. I’m also open to adding a social element like Stack Overflow to encourage each other in your Chinese learning. So you’re creating communities that encourage each other in language learning. I’ve seen a quite a few few startups working on that, but I don’t think anyone’s quite mastered it yet. Except for maybe Skritter, but those guys already left and are working on another startup.
How many people do you employ full time?
It varies. We work mostly with contractors, and we don’t have an office. Pleco has zero statutory employees at the moment. The number of people working full-time is probably three or four.
Is Pleco a full-time job for you?
It’s most definitely a full-time job for me. An additional benefit that I get, which I recognized early on as I started Pleco, was I knew that I would be a father someday and I knew that it would be nice to have my own business at that point. I could always be available and set whatever hours I wanted. The work from home thing is really a dream if you’re a parent, especially of young children. So now I work when I don’t have something else I need to do. It’s certainly 40 hours a week most of the time, sometimes 50 or 60 hours a week.
Is it hard to work on Pleco since you’re based in New York, where you can’t really ‘live the language?’
Not really. There probably are a few things I’m missing out on. My knowledge of Chinese slang is probably not as good as it could be, but I’ve got some people doing that for me. I guess I’d say that it’s helpful to be a part of the Chinese language learning community in China, but at the same time, it’s good to have a little distance from it. It helps me look at some of the problems with more of a critical eye. I feel like with software, you get enough usage data from emails – users give you feedback and complaints and compliments. Frankly, we have more customers that are students in the US or outside of China than customers who are from China. The stereotypical user of Pleco is the Westerner who’s living in China studying Chinese.
Some people say Chinese has become easier to learn thanks to the web and mobile tech. Do you agree or disagree?
I very much agree. It’s gotten easier because you can spend less time on grunt work. It’s much easier to find language learning materials, and it’s much easier to find language learning buddies. Of course, you can’t change much about human memory. For all the creative tools we try to engineer, learning Chinese is still hard work and you have to memorize a lot. No one’s really come up with a better way to learn grammar as far as i know, or make your tone pronunciation better. The only way to learn some of that stuff is to spend a lot of time around native speakers. So an awful lot of the problems are as hard as they ever were – the only thing is that we’ve gotten rid of some annoyances.Editing by Steven Millward