While computers have shown incredible progress in utility since inception, there is still a tremendous amount of potential to be realized. We’ve been saying for years that one of the keys to unlocking this potential is making computers easier to use. We’ve spent virtually all of computing history teaching people to use computers, when instead we should be teaching computers to work with people. It is with this tenet in mind that we awaited the debut of Siri with bated breath.

As most know, Siri is a new voice-based interface to Apple’s iPhone, and became available a few weeks ago with the introduction of the newest model, the iPhone4s. Apple fans and detractors alike were intrigued with Siri’s functionality and certain clearly intentional quirks.

It’s pretty clear to smart-aleck techies that one of the development exercises for Siri included getting a bunch of techies to ask a bunch of smart-aleck questions.

For example, if you ask Siri, "What’s the best smartphone?" Siri answers, "There are other phones?"

Humor aside, Siri does provide some much-needed utility. Texting while driving is illegal in just about every state in the union and most Third World countries, but Siri provides a perfectly legal way to text while driving. Simply tell Siri to relay a message, e.g., "Tell Mom I’ll be late for dinner," and Siri figures out Mom’s cell phone number, translates the message into text and sends it over.

While there are many more examples of Siri’s utility, its limitations are exposed as well. Following our previous example, if you say, "Tell Mom I said I’d be late for dinner," Siri will text Mom with "I said I’d be late for dinner." Clearly this is not the exact message you want to convey, and, in fact, could be easily be interpreted to show irritation or agitation.

While this is a minor example, it points to a larger issue. Siri, like many other technologies, is following a strict algorithm to respond to the user, instead of trying to understand the user and act accordingly. Again, the person must be taught to work with the computer (don’t say "I said"), as opposed to teaching the computer to work with the person.

Technologies such as Siri and IBM’s "Jeopardy!" whiz Watson still have such limitations. The good news is that these high-profile products are indicative of tremendous efforts to teach computers to work with people. These are definitely steps in the right direction and should not be pooh-poohed despite some obvious shortcomings. Both developers and users will learn from such efforts and build upon them.

John Agsalud is an IT expert with more than 20 years of information technology experience. Reach him at johnagsalud@yahoo.com.