Information architecture

On another mailing list, I heard about a T-Mobile Web page that would tell you what kind of signal you are likely to get at a given street address in the U.S. Because U.S. Cellular offers nothing but bottom of the bin cellphones (they are not big enough to get deals to sell phones like the Treo 650), I thought I'd check T-Mobile.

The company has never had coverage in Blacksburg, but I thought I'd try again, since I have not checked with them lately. So I went to the page, typed in my street address, city, state, and zip code, pressed the button, and voila.

I got back a message saying "Input zip code is invalid."

Wow. That's interesting. Either T-Mobile is trying to tell me I live in an "invalid" place, or there is a bug in their code, or it's a really awful way of saying they have no coverage in my area.

Pick any one of those three choices, and you get to the same conclusion--somebody screwed up, either by not testing it adequately and/or by failing utterly to do a basic software ergonomics review to make sure the "error" messages made sense.

My guess: This little app was outsourced to a software shop in India, which did a bang up job of banging out the code cheaply and producing a slick little application. But you get what you pay for. Outsourced contractors rarely care much about little details like this; they are under the gun to get the work done quickly and cheaply for the client, and so they don't have the luxury of dotting the i's and crossing the t's.

At the same time that we see big companies getting bigger by relentless costcutting and globalization of production, why is there a parallel rise in small, entreprenuerial enterprises? Because the small entrepreneur has more skin in the game--he or she has to produce high quality stuff to be competitive. A two hundred person coding shop in India just has to get the next job. It's not either/or here....both have their place, and both can provide useful services.

GovTech has an article on Missouri's new CIO (Chief Information Officer), who was given the daunting task of improving state government IT services, in part by consolidating 16 separate IT fiefdoms. IT folks are notoriously resistant to service aggregation, because it usually means smaller staffs and smaller budgets. Some IT folks like big, complicated, hard to use systems because it justifies big IT staffs and budgets.

The FBI wants to increase the cost of Voice over IP. The VoIP news article has a set of excellent questions that someone ought to be asking the FBI as they seek to extend existing wiretap requirements to VoIP companies. Not only will it increase the cost of commercial VoIP software by requiring those firms to install wiretap backdoors in their systems, the whole exercise is absurd. Here's why.

• As VoIP News asks, why would criminals use a commercial VoIP offering that was known to have a wiretap backdoor when they could just as easily use their own secure VoIP software. Dozens of VoIP software products are completely free and can be downloaded and installed easily.
• Some VoIP providers are located outside the U.S., beyond the jurisdiction of the FBI. Why would anyone use a higher-priced U.S. service if a less expensive offshore service with equivalent voice quality is available? Why would international drug rings ever use U.S. services if the FBI has their finger in them? One effect of FBI regulation will be to drive the entire VoIP business out of the United States.
• Wiretapping only the VoIP data streams of suspected criminals is, well, just dumb. If I were trying to investigate criminal behavior, I'd want to capture their entire data stream. And the FBI can do this now, just by going to the criminal's ISP with a court order and getting the ISP to re-transmit to the FBI every data packet coming from the criminal. This is trivial to do, does not require expensive new software, and is much more likely to provide useful information, since you'd also see email, Web sites, chat, IRC, and any other communications, along with VoIP conversations.

So what's really going on? Occam's Razor may be useful here (the simplest explanation is probably the correct one). Recall that this is the same FBI that just spend $170 million of our tax dollars on a "Virtual Case File" system that does not work. In other words, the FBI has neither good in-house technology advice nor do they seem capable of buying it. Like many other Federal government agencies, when the FBI wants technology, they run to the beltway bandits--the big consulting firms that inhabit the D.C. area, who have a built in conflict of interest when asked by those same agencies to both design and build systems.