Anomaly in date difference calculation between Chromey and WolframAlpha.

[GoogleCodeExporter]

This might be a matter of convention but how are these results reconciled 
between WolframAlpha and Chromey? 

WolframAlpha query "7 jun 2010":
"Time difference from today (Friday, August 6, 2010): 1 month  29 days ago"
"60 days ago"

Chromey query "6 aug 2010 - 7 jun 2010":
"1 month 30 days since June 7"

Chromey query "6 aug 2010 - 7 jun 2010 in days":
"59.42 days"

It seems the WolframAlpha answers make more sense. Calculating manually gives 
me "60 days ago" which gives "1 month 29 days" if a month is defined as having 
31 days. 

If Chromey means 31 days by "month" then it is saying 61 days have elapsed, yet 
it gives "59.42" days. But if Chromey defines a month here as 30 days (the only 
other rational choice), then it should answer "1 month 29 days" given that it 
reckons a figure of "59.42" days ago, because 59.42 - 30 = 29.42 days, or 29 
days to the closest day. And if Chromey is taking the time of day into account, 
then what rational and consistent convention used could result in an answer of 
"59.42" days?

I made these queries on 6 August, 2010, in South Africa (GMT +2:00) around noon 
local time. The queries need to be run to a date one day after two months into 
the past. My OS is Vista. Chrome version 5.0.375.125. I had the Chromey 
extension installed for a few minutes to try out, but uninstalled it before 
trying these queries.

Original issue reported on code.google.com by stefan.c...@gmail.com on 6 Aug 2010 at 10:14

[GoogleCodeExporter]

The 59.42 comes straight from WolframAlpha:
http://m.wolframalpha.com/input/?i=6%20aug%202010%20-%207%20jun%202010%20in%20da
ys

Chromey doesn't define "month". In fact, Chromey defines almost nothing; it 
forwards queries to Google and WolframAlpha.

The missing time seems to be WolframAlpha bug.

Original comment by iisi50...@gmail.com on 20 Sep 2010 at 3:16

• Added labels: Component-Logic
• Removed labels: ****