There's more to Google's mobile phone plans than meets the eye, as this diagram from the company's June 2005 patent application demonstrates.
The wraps are coming off Google's cell phone skunk works -- now it's time for the company to take a dose of its own mobile medicine and improve its existing mobile software

Kids say the darnedest things.

After noticing that I was writing a story about a stealth mobile phone development effort at Google earlier this week, my inquisitive high-schooler posed a stumper: “Why would Google even make a cell phone?,” he asked. “Aren’t they a search engine company?”

After pondering for a moment, the best answer I could come up with was, “Because they can.” After all, why would a successful search engine company assemble a group of engineers to design a new cell phone? What could Google possibly hope to gain, other than a foothold in yet another market? Is it just an ego-driven land grab by the gargantuan company’s overly acquisitive executive team?

More importantly, it wasn’t immediately apparent what value Google could possibly add to its mobile device that would set it apart from other innovative entries, such as Apple’s upcoming iPhone. Does the world really need a Google cell phone?

The real answer came a few days later, when the proverbial other shoe finally dropped. A recently granted Google patent  has come to light that makes the search giant’s intentions crystal clear. It all revolves around search capability. Go figure.

The patent, submitted in June 2005, specifies a method of “method of providing text entry assistance data” that relies on “predictive textual outcomes relates to search requests made by a plurality of remote searchers.” In retrospect, the answer to the questions posed above couldn’t be more obvious. If you could boil Google down to one core concept that differentiates its search abilities from its competitors, it would be the ability to aggregate hundreds, thousands or even millions of Web links to deliver the most likely search results based on popularity. Google calls this technology PageRank, and touts it as “the heart of our software.”

Now imagine the same concept being applied to text entry and Web searching on mobile phones. Anyone who has tried to enter text into a mobile device, whether to send a message or perform a search, is well aware of the limitations. To compensate for constraints such as a limited number of keys that represent multiple characters, a variety of schemes have been employed. The best known is a predictive system called T9, which uses a word database, or dictionary, to try to guess which word you’re trying to enter and complete it automatically. This autocomplete function is a lifesaver, but not without its faults. For example, it often guesses incorrectly, and in most cases, it never learns from its mistakes.

Now imagine how much better T9 technology would be if it paid attention to your preferences and usage history, and had access to the usage history of thousands of millions of other thumb-typers by tracking their mobile text entry habits? I know red flags are suddenly being hoisted all over the hypersensitive security communities, but let’s just assume for one moment that this tracking and data aggregation was being used for the purposes of good only, and not for evil. Accuracy of such a system would be potentially through the roof!

Well, that’s apparently  just what our friends at Google are proposing. In fact, their patent is based on the following elements:

  • Recording and analyzing user-specific textual input and search requests
  • Recording  and analyzing search requests made by a plurality of remote searchers
  • Using the information to generate a data dictionary for predicting text input
  • Providing that dictionary data to a remote device

Now let’s take that a step further. What if the data Google provided also took the phone’s location into account? Now text input and search results potentially become even more accurate, delivering different predictive results if the user is in Chicago, versus Los Angeles or Vancouver BC. This capability it also included in the Google patent.

Now I only have one question left. What is Google waiting for? This technology could be rolled out immediately within Google’s server-based mobile applications, such as Gmail. In fact, one of the biggest drawbacks I’ve found to actually using Gmail as it exists today is the inability to input messages easily. While users are offered a T9 word guessing option, it ungraciously insists on capitalizing every word, A Trait Which Most People Find Annoying.

So my suggestion to Google is don’t wait until you’ve got a hardware device of yours. Try a dose of Google’s homegrown predictive text medicine now. Your existing mobile users will thank you.

"I'd be pissed too, but you didn't have to go all Minority Report on his ass!" -- Jon Stewart on police raiding Gizmodo editor Jason Chen's home
Related Articles

Most Popular Articles

Copyright 2018 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki