So-called data phones have not sold well so far. I do believe in the integration of telephony and computing, but data phones will not take off as long as they are designed from the wrong conceptual model. So far, all these devices have been designed as telephones with a data add-on. They'd probably be more usable, and more successful, if they were designed as as computers with voice communications capability added on.
The difference lies in the user interface: telephone user interfaces are horrible, which is why nobody can figure out how to use services like call waiting, much less how to forward a call to another extension in the office! Computer interfaces aren't perfect, but the usability and the design of multiple features is better done based on computer thinking. Users need an integrated user interface rather than something that is half-telephone and half-kludge.
This is not to say that Windows CE is the solution: it is not appropriate to take a user interface that was optimized for large-screen desktop devices and use it for small-screen handhelds. The smaller the device, the more strict the requirements to optimize the interface for its specific characteristics.
Some people claim that the telephone is an example of perfect usability that should be emulated by software designers. After all, it's easy:
you pick up the handset
punch in the number
you are connected
If only it were that easy in the real world. Of the three steps,
only picking up the handset is truly easy
. Turning on the device and "logging in" to your account are both accomplished by the simple action of picking up the handset. There is no "boot time", and the dial tone is always there. Computers (and in particular the Web) can definitely learn something from the uptime requirements of the phone system. The Internet should supply users with "WebTone" at the same level of reliability as the telephone supplies dial-tone.
Let's debunk the myth that punching in a number is an easy-to-use user interface that should be emulated. First, these numbers are actually hard to learn and remember.
Quick, what's the number of your dentist?
Second, they are hard to type, and there is no forgiveness if you mistype a digit: nothing to do but hang up and start over. To make a long-distance call from a typical office in the United States requires the user to type in twelve digits: this is quite cumbersome and takes a long time to do. International calls are even harder.
The real usability problems of the telephone show up when we do a task analysis. What does the user really want to accomplish? In most cases, you
want to talk to a specific person
. To do so, you have to find that person in the telephone directory (or another list of phone numbers) and then dial the number. Who wants to talk to a
It would be better to be able to search a database and click on the person's name or photo to get connected, something a computer interface does well. Furthermore, most people have many telephone numbers: office, home, cellular, fax, secretary, etc. Sometimes several of these numbers have their own voice mail system attached (each having its own inconsistent set of features and commands).
Finding out what number to call should not be a matter of guessing where the person is, nor should you be forced to try all the alternate numbers until one works. A computerized communications system would know where the person is and what device currently is the preferred way of reaching him or her.
An integrated communications device may first try a voice call, but if the person is not available, then the fall-back should not be voicemail but the person's email if available. Similarly, when voice mail is used, you shouldn't be forced to access it by sitting through a long, linear recital of recorded messages; scrolling through a visual listing of messages would be more efficient.
Another benefit of an integrated communications system is the opportunity for the recipient to specify when he or she is willing to take calls from whom. Currently, anybody who can find your telephone number has a license to bother you (and many telemarketers do so without shame). It is easy to imagine a communications system where the two parties are represented by software agents that negotiate whether or not the call should be allowed to go through, depending on urgency and the caller's level of prominence. For example, you could give out encrypted tokens to your family and closest colleagues that would allow their calls to go through at all times, whereas other callers might only be allowed at certain times or if they pay you, say, $10. The same system could screen both voice calls and email and faxes, though you might be willing to accept unsolicited email for a lower fee (say, $1) because it is less intrusive than a telephone call.
Much more can be done once you start thinking of all communications media in an integrated manner. It is all a matter of fundamental perspective in design, and so far, the telephone industry is still not getting it: voice is just another datatype.