>city/state).
>
>http://weather.boygenius.com/oklahoma-miami_beach.xml
><http://weather.boygenius.com/oklahoma-miami_beach.xml>
>
>I promise you there is no Miami Beach, Oklahoma....I would've already gone
>there, oh well.
Late chiming in, perhaps this is still relevant to someone.
The orphan cities are a consequence of the NOAA date that I'm parsing to generate the XML files. The NOAA data is stored in files that look like this:
SOUTHWEST OKLAHOMA
CITY SKY/WX TMP DP RH WIND PRES REMARKS
CLINTON CLOUDY 21 8 57 N3 30.48F
HOBART CLOUDY 23 10 57 NE9 30.49F WCI 13
The file is called something like 'weather-roundup-ok.txt'. All my script does is say "ok, we're looking at ok, that's Oklahoma, so take these lines and put them in files called 'oklahoma-clinton.xml' and 'oklahoma-hobart.xml'.
The problem with this is that the NOAA isn't really strict about state boundaries in their data. New York City, for example, appears in the New York file, the Pennsylvania file, the New Jersey file, and (oddly enough) the Massachusetts and Maine files.
Well, fugly data calls for fugly parsing, so I have this subroutine that says "If you see New York City, Massachuettts, it's really talking about New York City, New York." This works fine, except that it requires me to have a line in my script for every NOAA geographical generalization. Every once and a while, a new one pops up, and you get things like oklahoma-miami_beach.xml.
This is noted in the changelog on the site:
"20030902 - I think I have all of the orphan cities taken care of. If you notice any that I've missed, please tell me."
I'll note that you didn't bother to tell me. Thanks.
Anyhoo, the file you listed above would have contained the correct weather information for Miami Beach, it would have just had the wrong state name. I don't consider this to be a big deal, because if you want the weather for Miami Beach, the place to look is florida-miami_beach.xml.
Todd
[Todays Threads] [This Message] [Subscription] [Fast Unsubscribe] [User Settings]
