Well I just tried the above link and the times were jumping. I did notice though that when the times are erratic that the location is N/A rather than Paros Community Airport, Greece.
This <lsup>9/12/06 12:12 AM Local Time</lsup>
<obst>N/A</obst>
rather than <lsup>9/11/06 5:00 PM Local Time</lsup>
<obst>Paros Community Airport, Greece</obst>
It could be a timing issue, how frequently do you have the WW set for? Or it could be a weather.com problem with the airport’s server. But WW could test for this and display an error msg when the weather.com data has not been updated rather than updating the forecasts with incomplete data. Yes?
Ok so we found the bug… Yes it gives a N/A on the Paros Airport and it displays all kinds of different erroneous data on the Observed time…
The bug could be in the weather.com…
I was hitting update and the observed kept changing and changing but all the values were constant… 21 degrees and 6 beauforts North winds… I changed the active city and then back to sifnos and the winds were 4 beauforts now, which is probably wrong again…
So write to weather.com and tell them that the Today weather data is missing for Sifnos, Greece when you pull it up on Weather.com. (The other tabs all work!)
You can leave out telling them that it’s screwed up in WW also. WW is not their product and they offer no support for it, as such they will have no incentive to resolve WW’s problem. But they will look into resolving the weather.com problem and hopefully the fix for one will be the fix for the other. :icon_smile:
I think Mike should look into having a unreleased version of WW that checks for valid data before updating the forecast’s html files that you can use.
In the interim you should look for another nearby weather location. One that works.
I can’t write to them since I couldn’t fint an actual email address to mail them… But I wouldn’t be the right person also since you guys are more fluent in this “language”…
Anyway, I changed the active city to Paros (next island) and it seems more stable so far…
Mike… I’m definatelly very serious about the wind request that I made in a previous post about the 6 gusting to 8 etc… Is there a way that you can fix this for me?
Wow!! :shock: The problem’s back. Hopefully you wrote to them and informed them of such. The problem will get fixed a lot sooner if they are made aware of it.
Any changes to WW planned to properly handle this type of situation.
I believe they’ve always had this “problem”. It seems to occur much more often with international cities. Perhaps their international data comes from a different source.
I doubt it… it’ll probably just give them more e-mail to not respond to. I have never heard back from them on anything.
Nope. It looks like this is something they need to fix on their end.
While it’s prudent for any program accepting input from an outside source to check it’s validity before processing it, WW doesn’t have to check the data it simply has to display the data it reads from weather.com. When the Observed times are jumping the xml data has N/A for most of it’s values. Yet WW continues to display numerical values leading viewers to believe they are actual numbers.
There is definitely a bug somewhere on weather.com’s end. It doesn’t make sense for me to add a hack to Weather Watcher to work around their bug though.
Sorry, but displaying data that doesn’t exist is misleading, regardless of the source or reason.
Professional apps edit the input and display error msgs when the input is wrong. WW doesn’t have to do that, it simply has to display the data Weather.com distributes. And at this point WW is not doing that. When the xml data fields contain the characters “N/A” WW is substituting numbers from some other source and making the WW forecasts appear to be valid when they are not.
It’s not even possible to check for the problem in the templates because WW.exe is hiding the error.
The link at the bottom of the WW forecasts should at least be changed to read “View may not agree with the actual forecast displayed at weather.com?”. :icon_smile:
I was going to include a portion of the xml data for Sifnos showing the N/A values but weather.com has fixed the forecasts, again.
As a matter of fact, I do! Sifnos is doing it again. I wonder if this happens when they do a nighly backup, or someone shuts the power off when they leave for the night.
I suspect WW must be testing the xml data for numerics and if not numeric skipping the Registry update and then using the old Registry data in the templates. Yes?
We appreciate your alerting us to a potential problem with our Web site or one
of our products. We’ll look into the issue immediately and make any necessary
changes.
We apologize for any inconvenience you may have experienced. We hope you will
return soon.
The Weather Channel Interactive
Brittany
I think I last wrote to them on Monday. It will be interesting to see what the data looks like tonight.
Don’t get too excited. At 16:30PM EDT Sifnos is back to wild Observed times and N/As for Current data.
If you can get the good people of Sifnos to write to weather.com, using the link noted above, it may speed up the fix. Remember, the squeaky wheel gets the oil. :icon_smile:
