Category Archives: UX

Have you tried restarting the network, sir?

For most people who have fast (broadband) connexions to the Internet, and have called their ISP’s tech support, that line is quite ubiquitous. However, I have to report that doing it actually solved something for me!

Continue reading

A new security question

We all know those “security questions” that websites ask to verify user’s identities, sometimes called shared secrets. The ones like your mother’s maiden name and favorite high school pet’s phone number.

I propose a new one for technical people: What was the name (SSID) of the first wireless network you connected to?

At the beep, enter your language

If you live in a country or area that is at all non-monolingual, you will encounter telephone trees where the first prompt requests you specify your language. For example, in the United States, it is usually Spanish.

Unfortunately this does not scale well. If you have to deal with even, say, the UN languages, that’s 6 possibilities. Further, they are rarely standardized, so for one firm, you key in “6” to get Spanish. For another it’s “2”, etc.

I propose this:

  1. A universally understood tone or sequence of tones that means “specify your language” (SYL). These would be tones that a computer could recognize, like SITs. In this way, a person could specify to their phone/phone company what their language was and have them automatically reply.
  2. A universally standardized mapping of languages and dialects to numbers. For example, en-US = 1033 (Microsoft LCID).

This would result in the following pass:

  1. Caller dials some number with a phone tree.
  2. Called party PBX picks up.
  3. Called party PBX plays SYL SIT.
  4. Caller (or caller’s phone/phone company) recognizes the SIT.
  5. Caller (or caller’s phone/phone company) responds with language code
    1. This would include a termination character, like # or *.
  6. Called party PBX connects the caller to the phone tree or operator of that language.

Of course, even the UN isn’t going to maintain an operator for every possible language, so in those cases, a fail-gracefully routing tree would be set up so that the nearest neighbor language would be selected instead. As an example, if en-GB (2057) wasn’t supported, but en-US was (1033), the call would be routed there. Alternately, a message could be prerecorded in that language, telling the called party that their language wasn’t supported.

What do you think?