No Canadian weather? That sucks, Asus!

I just got a ASUS O!Play HDP-R1. It’s one of the current crop of media player boxes, like the WDTV. I suspect they’re all the same MIPS hardware/Linux firmware inside. (Hey, you can telnet into it! Whee! Or something.)

At first, I couldn’t get it to work with my network share, but after a firmware upgrade, all is good. The new firmware offers web content, including weather. So I hopped along, and tried to access Toronto weather:

Err, some of us live outside Asia, Europe and the United States, y’know …

If Burlington = 95, Toronto = ??

143; if you’re Environment Canada’s Weather Office, that is.

Completely against the concept that Cool URIs don’t change, the Weather Office appear to have arbitrarily changed the URLs for their 5 day forecasts. In the old days, if you knew the local airport code, you could find the weather report, for it was at (f’rinstance): http://www.weatheroffice.ec.gc.ca/forecast/city_e.html?yyz. (And yeah, I grew up with siblings obsessed with these guys, so I’ve known Toronto’s airport code of old.)

Now the same page is at http://www.weatheroffice.ec.gc.ca/city/pages/on-143_metric_e.html, and for Burlington, it’s …/on-95_metric_e.html. It doesn’t make sense, does it?

Canada has weather that can kill you. Somewhere, someone will have scripted a page that scrapes the Weather Office’s data, and so somewhere I’m sure there will be a weather report that’s not updating. I can see no good reason for this change; I’ll see if Environment Canada has one.

It’s not as cold as you’d think

I use Gnome Weather Report, an applet that shows the local temperature and weather conditions on my desktop. For the last few days, it’s been showing something really weird: celsius.png. It’s nothing like -17°C here; it’s nearer 0°C, according to Environment Canada.

Things become clearer when you change the view to Fahrenheit view: fahrenheit.png. It’s clear that the sensor or protocol is broken, but is being mis-interpreted as a zero signal.

As an avid RISKS reader, I know that confusing zero and null values is pretty much unforgivable. I’ve wired up enough 4-20mA current loop instruments to know that having a zero-value signal being the same as a no signal value is bad.

But there’s no real risk here. I mean, I could always go outside and find that it’s not 17°C. You don’t need a weatherman, as Bob said.