Translation Guy Blog
Word Lens is an iPhone app that will translate Spanish signs through the iPhone camera, as posted here a few weeks (or is it months?) ago.
A few days after that post, the CEO, John DeWeese, sent me a message on LinkedIn:
“I enjoyed your blog entry on Word Lens, but my favorite part of it prevented me from replying in comment form—after all, I’m no 50 cent troop! There’s just the two of us over here at Quest Visual, former video game programmers who wanted to take a stab at an important problem. We’re not paid to hock party lines, but I do want to set people’s expectations.
“I had heard about 1-800-translate a few years ago. I have great respect for what you guys do, and we are truly newbies in comparison. But we both know the value of this problem, so it’s worth a look from our wildly different perspectives.
“We got caught in a viral video storm. Meh. It’s more important for me to go through the looking glass, darkly first if need be, to figure out how to do it right. Within a month, the interwebs have told us that:
1. Our translations stink
2. It’s still worth doing
“So we’re going with it. Gotta put the idea out in practical form, if the business is to survive.”
I was really impressed by John’s candor, and was anxious to talk with him, since he would certainly be dropping the F-bomb and I would get all these great quotes. As it turns out, all the F-bombs were mine (hardly surprising) and I forgot to record the conversation, so no quotes either. But I learned a lot about translation apps on iPhones, and I hope to enlist the comments of readers so that John and I can learn more.
Word Lens was conceived as a tool that could be used by tourists to find their way around foreign cities. So John’s idea was that the phone had to be able to function in real-time on- or off- line. Latency is a show stopper, which, if you think about it, is the big problem with translation in general. The reason we settle for the Google Translation app is because it’s right now, accuracy be damned. The reason doctors won’t use my telephone interpreters to save patients’ lives. Time is short.
So for an application that magically translates what it sees, every millisecond counts, so coding has to be short and sweet and has to run in the tiny computer residing in each Smartphone. John and his partner did all their coding from scratch, using none of the optical character recognition or translation engines publically available to make it all fit.
So the translation sucks. So what? Free translation is supposed to suck. The problem is when you start charging for it. Getting pinged in the Apple App Store is a Rubicon few free users will cross, as I can attest from personal experience.
John says, “We captured the imagination of the whole idea of augmented reality, but we want it to be a tool that is actually useful and handy.” So is it? It’s impossible to tell from the reviews on the app store, which are about as useful as YouTube comments in gaining insight into the user experience. “This app sucks because it isn’t free!” and stuff like that, so unlike the erudite seekers-of-knowledge who post here.
For augmented reality enthusiasts, this is a big deal. But since I find un-augmented reality pretty overwhelming, I’m sceptical about adding more input. Has anyone found this app useful beyond the “gee-wiz” factor? I’d love to hear your war-stories.