Ever since developers got their hands on Google Glass earlier this year, software coders have clamored for greater access to the programming internals of the controversial headset. Google accommodated them this week -- albeit to the sound of muted applause.
To be sure, Google's Glass Development Kit (GDK) does fill in a key puzzle piece that had been missing from Google Glass. Yet many developers are worried. They say that in the absence of more leadership or more access from Google, they're being asked to figure out the final picture on their own -- knowing that it might change by the time Glass gets mass produced.
The Glass Development Kit preview released by Google opens up many of the options that had been absent from the developer's toolbox. Previously, developers had only been able to code for Glass' limited Mirror API.
- Joy! Google Glass app tells you how other people are feeling
- Corning exec slams sapphire -- rumored for Apple device
- Google Glass updates slow down ahead of KitKat upgrade
- The definitive guide to not being a Glasshole
- The two things that need to happen before wearable tech goes mainstream: Google and Apple
Google has relied on third-party developers who own the $1,500 headsets to further app development, while internal development at the company has focused on making sure that the new software platform functions properly. When Google first announced Glass, the hope was that a vibrant development community would emerge and create the kinds of consumer applications which would extend Google Glass' appeal beyond the technophile crowd and into the mainstream.
Given that Glass represents a major shift in wearable computing from the nerdy realm of adventurous hackers to the common consumer marketplace, Google has been cautious about giving developers too much access to the hardware too soon.
Many but not all doors open for Glass developers
The Mirror API access was a compromise that encouraged developers to build for Glass when it arrived on their doorsteps last spring, but without giving them too much power. The GDK, which Google is quick to caution is a "preview" and not the full GDK that eventually will be distributed, gives developers access to many Glass features that had been walled off -- but not all of them.
"Now all that's remaining is for Google to build an actual app store, and for developers to build better apps" with the GDK, said Jonathan Gottfried, a developer evangelist at Twilio and developer who built early Twitter apps for Glass.
"I wouldn't be surprised if they went with the Google Play Store," he said.
Brandyn White, a 27-year-old Glass developer and self-described lifelong computer hacker who founded a consulting company that specializes in how computers interpret the world through camera lenses called Dapper Vision, has been working with Google and on his own to build the kind of better apps that Gottfried described.
--Brandyn White, Google Glass developer
White has focused his attention on how Glass can be used to help the visually impaired.
"For me, the most important thing is context. Take a grocery list," said White, who's also earning his Ph.D. in Computer Science at the University of Maryland. "Glass should recognize things at the supermarket," telling you when it sees something on your list without having to actually show the list to you.
"You want it to be a non-annoying friend; you want it to add value," he said. Glass needs more of those kind of apps, he said.
White cited apps like the translation app Word Lens pictured above, which are essentially interface-free interactions filtered from the real world through Glass to the person wearing it. Like its Android and iOS siblings, Word Lens on Glass replaces large font words on signs with words in your language. But it can't translate smaller print, even though its smartphone siblings can, because the Google Glass camera doesn't zoom in enough.
It's likely that current Google Glass hardware is more of a prototype than what consumers will get sometime in 2014, when Google has said Glass will be available to purchase. Think of it as the original Chromebook prototype, or the Nexus line of Android devices: It's a hardware guideline for the final product.
Nevertheless, White said that the hardware is perfectly usable in its current state. "The hardware team has everything figured out," he said. "The software is so much harder. [The Glass software team] has to think long-term, and legacy with Android."
White said that "80 percent" of Glass code is taken straight from Android.
Glass developers left to fend for themselves
He added that the GDK is not inadequate for developers, but that it could offer so much more. That, White said, is where Glass developers are taking the lead over Google's own team.
"The GDK is very small, it's very useful, but I think a ton of features got left on the floor," said White.
One of those restrictions appears to have been enacted to keep privacy advocates at bay. Google does not allow or encourage facial-recognition apps, severely limiting science research, said Vivienne Ming, the chief scientist at tech employment firm Gild, and a visiting scholar at the Redwood Center for Theoretical Neuroscience at the University of California at Berkeley.
"What I'm most fundamentally interested in is this idea of maximizing human potential," she said. "We could do expression recognition, and use it to teach autistic children how to recognize expressions." Another helpful scenario she described for facial recognition would be to help Alzheimer's patients remember people that they know they ought to recognize, but have forgotten.
That doesn't mean that facial-recognition apps for Glass won't be available. You just won't be able to install them through the MyGlass catalog until Google changes its policy. But since Glass is based on Android, and Android has built a lot of its success off of having multiple app stores, it stands to reason that there will be more options for getting Glass apps than the official Google distribution point.
Not waiting for Google to lead
Even now in these formative Glass days, there is already at least one other option besides MyGlass. White and a partner have built Weariverse.com as a place where Glass hackers and owners can share and install scripts directly to Glass.
Another big piece of the puzzle will be how Google decides to implement app sales. Currently, there's no viable business model that Glass app makers can rely on, said Winnie Tong, who built the photo filter app Glasstagram last summer, but has since stopped developing for Glass because of her day job.
"Right now there are no viable business models, which makes starting a startup very difficult," she said.
Geographic tagging in photos is another potential problem for Glass, she said. She attempted to write an app that uses real-time GPS coordinates and photos to help you remember where you left your car in a parking lot, but stumbled on the up-to-10-minute GPS location update delay built into Glass.
The GDK lets app developers tap into real-time location, she said, "except that geotags on photos still have an up-to-10-minute delay."
There are other issues that have yet to be resolved on the hardware side. Complaints about battery life are common among developers, and Glass accessories such as $85 mono earbuds and eye shields are in the works. How Google plans to integrate prescription lenses with Glass is another issue that the company is working on.
Google and its army of Glass developers have made impressive leaps in creating a Glass ecosystem of apps and real-world use cases out of what had previously been left to the realm of science fiction, but there's yet to be a single killer Google Glass app.
That may never happen, said White, because of Glass' wearable nature. "As the device gets more personal, it's less about killer apps, and more about the personally awesome app," he said.
The difference between killer app and personally awesome app could be as simple as Word Lens translating written words you see as you read them. The puzzle pieces have been dumped on the table, but Google's leaving it up to developers more than ever before to complete the picture.
Corrected, Tuesday, Nov. 26, at 6:50 p.m. CNET had misidentified Vivienne Ming in the photo caption.