I'm writing this in an Amsterdam apartment we rented for the weekend through AirBNB. The main reason we chose this place wasn't comfort or convenience. It was connectivity. This apartment was relatively cheap (about a quarter or a third of the price of a three-star hotel), but reports said the Internet connection was good.

And so it is. Speedtest.net tells me I'm getting 40Mb/s downstream and 2.5Mb/s up. That's a bit lopsided for my tastes (I upload a lot of stuff and back up in the cloud), but it's a heck of a lot better than what most of the hotels provide, which on the whole is awful. (At least in the US and in Europe, which is where I do nearly all my traveling.)

But when we leave the apartment here, we become data-poor. And that's by intent of the mobile phone companies that provide the only easily available source of Internet connectivity.

Open Wi-Fi access points, once plentiful, are now rare as four-leaf clovers and far more lucky to find. (Our many reports in Linux Journal on open Wi-Fi, published from 2002–2004, look like paradise compared with today.) That puts you at the mercy of cell-phone companies. Most (or all) of those in other countries would rather not sell data usage to short-term visitors, so your only choice is using your domestic phone.

Ours are AT&T phones. So let's look at what AT&T charges, both within the US and outside the country. Note that I'm doing this as a customer trying to figure things out, not as a math whiz looking for problem challenge. But that's what we have here, and if I fail, please correct me. However, bear in mind that the failing is not mine alone, but a feature of AT&T marketing.

For US customers, AT&T has three plans (http://www.att.com/shop/wireless/data-plans.html#fbid=6jmPMqgegnI):

• $20/month for 300MB ($.067 per MB, $.0067 per KB), then $20 for each additional 300MB.

• $30/month for 3GB ($.01 per MB, $.001 per KB), then $10 for each additional 1GB.

• $50/month for 5GB ($.01 per MB), then $10 for each additional 1GB.

So here's what you need to figure out before you choose one:

1) For both the $30 and the $50 plan, the rate is $10 per GB, which is $1 per 100MB and $.01 per KB.

2) The per-MB (or KB) rate in the $20 plan is 6.7x the rate for the other two plans, and "saves" money only if you stay under 300MB for the month, which isn't much data if you're an active user. As soon as you go over, you get charged for another 300MB, or $40 total for the month—meaning you're better off with the $30 plan. That is, if you know for sure that you'll use more than 300MB.

3) If you pay $30 for 3GB and use 5GB, or 6GB, or 7GB or 10GB, you pay no more than the person who uses the same quantity and pays $50 for 5GB. Yet the 5GB customer using only 3GB still pays for 5GB. So there's no reason at all to go for the $50 plan, even if you use 10GB/month. You're still paying $10 per GB.

4) You still feel screwed if you pay $30 and use only 1GB, because you're still paying for 3GB.

But all of that is dirt cheap compared to AT&T's international data rates (http://www.wireless.att.com/learn/international/roaming/affordable-world-packages.jsp#data). Let's unpack those:

1) "Pay-per-use" is $0.015 per KB in Canada and $0.0195 per KB in "Rest of World" (aka RoW).

2) Since there's 1,000 KB in a MB, that's $15 for 1MB of data in Canada and $19.50 for 1MB everywhere else (actually, 140 countries). If you use 1GB, which is 1,000MB, that's $1,500 in Canada and $1,950 in RoW.

3) The above, therefore, shakes you down in to a plan. There are three to choose from:

• $30 for 100MB.

• $60 for 300MB.

• $120 for 800MB.

Translated to GB, that's $300 for 1GB at the 100MB and 300MB quantities, and a bit less at the 800MB level—and that's all for less than 1GB of usage. Even if you pay the top price of $300 for 800MB, you're paying $3,900 for the next 200MB, because you're in pay-per-use-ville above the 800MB level.

In other words, you pay $4,200 if you use 1GB while paying $300 for AT&T's top International data plan. (And never mind that AT&T doesn't tell you that's for both upstream and down, added together. I found that out when I called the company on the phone once to talk about it.)

So my wife and I each went with the $30 plan.

Now, since all this data is metered, how can we monitor it on our end?

With iPhones (which we're using here), we need to go into Settings→General→Usage→Cellular Usage after landing outside the US and then press Reset Statistics, erasing whatever record we had of use since the last reset. (The only memory we have of the statistics is a screenshot. This sets Cellular Network Data use at zero—presumably.)

When I did that, while on Wi-Fi here at the apartment, the meter immediately read 13KB sent and 56KB received. For what I had no idea, but already I felt screwed, because I wasn't on the mobile data network. (VodafoneNL, the phone tells me.)

When we walk around Amsterdam, we have to keep our Cellular Data and/or Data Roaming off, until we want to look at a map or pick up e-mail or some other thing that doesn't invite down a huge blast of data. And then, to monitor use, we have to drill back down that same directory path to see where we stand. My wife has done most of the roaming. In two days, she has received 9.5MB and sent 688KB. Is that for real? How can we tell? And why should we care?

I'll grant that the cost of connecting people is more than zero. But is the cost they're passing through that of data? Or are they just charging for sums of data because they have to charge for something, so why not?

I don't know, and I would welcome some answers, if readers have any. I know a lot of people in and around these businesses and still have not heard satisfying answers.

Meanwhile, it seems to me that there are matters of economic externality to consider. High data use costs and fear of bill shock cause negative externalities through lost business and diminished economic activity. Also, given that the Internet's cost-free base protocols comprise one huge positive externality generator, it should seem wise of mobile phone companies to participate in that economy, rather than restrict the whole thing to what little they can extract from customers they clearly enjoy frustrating.

A few minutes ago, a friend close to the business told me the problem is that the big mobile phone companies, at least in the US (that would be AT&T, Sprint, T-Mobile and Verizon) actually hate their customers. In fact, we take this for granted. In a Wall Street Journal essay of mine last summer (http://online.wsj.com/article/SB10000872396390444873204577535352521092154.html), the pull-quote they put in bold face was "Choosing among AT&T, Sprint, T-Mobile and Verizon for your new smartphone is like choosing where you'd like to live under house arrest." It was the line quoted most by others as well. But my friend took the hate thing a step further. He said treating customers like prisoners eliminates the main sensory path between a company and the marketplace, thus blinding the company not only to externalities of many kinds, but to opportunities as well.

So I take heart in another piece of news I picked up today: Mozilla's Boot to Gecko, or B2G (https://wiki.mozilla.org/B2G), now re-dubbed Firefox OS (https://developer.mozilla.org/en-US/docs/Mozilla/Firefox_OS), "uses a Linux kernel and boots into a Gecko-based runtime engine, which lets users run applications developed entirely using HTML, JavaScript, and other open Web application APIs".

In other words, it's smaller, lighter, simpler and less complicated to deploy than Android. Here are the components listed on the Firefox OS page:

• Gaia, the user interface, is "a Web application running atop the Firefox OS software stack".

• Gonk, the OS layer under Gaia, "consists of a Linux kernel and a hardware abstraction layer to which Gecko communicates".

• Gecko "is the layer of Firefox OS that provides the same open Web standards implementation used by Firefox and Thunderbird, as well as many other applications".

On the Booting to the Web page (https://wiki.mozilla.org/Booting_to_the_Web) in the Mozilla Wiki, it says this:

Mozilla believes that the Web can displace proprietary, single-vendor stacks for application development. To make open Web technologies a better basis for future applications on mobile and desktop alike, we need to keep pushing the envelope of the Web to include—and in places exceed—the capabilities of the competing stacks in question.

We also need a hill to take, in order to scope and focus our efforts. Recently we saw the pdf.js project expose small gaps that needed filling in order for "HTML5" to be a superset of PDF. We want to take a bigger step now, and find the gaps that keep Web developers from being able to build apps that are—in every way—the equals of native apps built for the iPhone, Android and WP7.

In July, TechWeek Europe reported a Firefox OS prototype being shown off by Telefónica, which is big in Europe and many Latin countries (http://www.techweekeurope.co.uk/news/telefonica-firefox-os-smartphone-prototype-85340). Here's the closing line from that piece: "The first handset based on the platform will be released in Brazil on Telefónica's Vivo brand in Brazil next year and will cost less than $100." Yay.

They're playing a value creation game, instead of a value subtraction one. It's a winning strategy in the long run. Let's help Mozilla make that run as short as possible.

Wi-fi image via Shutterstock.com.