Signal to Noise

I have been doing quite a bit of development for hand-held devices lately, mostly the iPhone and Windows Mobile. I took a brief foray into Blackberry and realized that I did not have enough high-quality booze to get over the learning curve (as close as you can come to a step function and still have rounded corners) and Byzantine documentation and API choices. I felt rather guilty at not being able to master the Blackberry and will probably give it another try when I have more time and fewer deadlines.

I came across a story about a developer who decided to write an application for the Blackberry and documented some of the ups and downs of his work. The article concentrates a lot on the aspects of writing an application that come after the development is “finished” (at least released) and the difficulty in converting the application into customer dollars. That part is interesting but the part that caught my interest was the first third of the article that deals with the development process and some of the difficulties he experienced. The part where he describes choosing the development environment hit very close to home as this was early in the process for me (and him) and it left me rather disheartened early in the process.

All BlackBerry phones are java based, and there are two SDK options for developing software on them: MIDP/CLDC and BlackBerry specific APIs. … Initially I leaned towards CLDC to maintain portability, but I eventually decided I’d rather take full advantage of the platform. I could always port it later if it came to that (if there is a later).

Great, now I just needed the RIM SDK. Turns out there are about 5 different versions, and the more features you get the fewer devices you can support. RIM has been around for a long time, and unfortunately (and unlike the iPhone) not all operating system revisions are available for all devices. The basic choices when I started PodTrapper were 4.2, 4.2.1, 4.3, 4.5, and the newly released 4.7.

He also hit a snag that I experienced (and the one that eventually broke my will), connecting to the network:

There are 10 different network transports available on BlackBerry: WiFi, Direct TCP, WAP, WAP2, BES/MDS, BIS, Unite, BES Serial Bypass, USB and Bluetooth. A lot of options for getting data in and out. What would be nice is to say:

Me: Give me an HTTP connection using least-cost routing.
BB: Here you go

What you get is:

Me: I need an HTTP connection, is Wifi available?
BB: No.
Me: Is BES available?
BB: No.
Me: Is BIS available?
BB: Yes.
BB: Ooops, that file is over the size limit for BIS.
Me: Is TCP available?
BB: Yes.
BB: Ooops, TCP looks available but it was blocked by the carrier’s firewall…..

Regdarless, he persevered and ended up with a decent application. I envy his dedication and results. Mine will have to wait until next time. I know it can be done now.

When my wife was taking some courses in nutrition at UPEI she would often bring home nuggets of wisdom that tended to cause us to eat food closer and closer to alfalfa. Generally, when I heard that she had picked up a tidbit from class I would shudder at the pending switch in foodstuffs. There is one occasion (well, more than one, but I am going to write about one now) when I enjoyed the nugget of wisdom. It went something like this

there is a study that has shown that men who wear black socks are more likely to have heart disease

This would be a good time to point out that I do not know if this study exists and if it does I do not have a reference to it. The point the professor was making is that some types of studies are “epidemiological studies” that attempt to correlate various factors of human populations in order to discover causal relationships between these factors. In other words, get a bunch of people and look at what is the same (or different for that matter) for that bunch of people and see if there is a reason for for the sameness (or difference).

If it turns out that men with black socks are more likely to have heart disease (which might be true) it does not mean that black socks cause a higher incidence of heart disease. There is a correlation but not a causal relationship. Another favourite of mine is the assumption that the crowing of a rooster causes the morning to arrive. The rooster always crows before daybreak and the day always comes. That is a slightly different argument that suggests timing implies a causal relationship (post hoc ergo prompter hoc) but I like it as an example, even if it does not prove anything about heart disease or epidemiological studies.

All of this would mean very little if it were not for the fact that I was catching up on XKCD and came across the above comic a couple of hours after I had a relatively long discussion with my aforementioned wife about High-Fructose Corn Syrup and the correlation (though potentially not causal relationship) with North American obesity. You may be jealous or relieved you do not have such conversations with your partner–I think you should be jealous, but that is just me.

Anyhow, I am still no fan of HFCS and I am not going to wear black socks. You have been warned.