Africa - data error

[Paul Eggert]

On 08/23/10 03:51, Jamie wrote:

> The International Atlas says that the city and country have observed
> a time zone of -1 (Central Europe Time) since 1960.  Before that the
> time zone was 0 (UT).

My copy of the International Atlas (6th edition, dated 2003)
also says this:

 Before 1 Jan 1912 LMT
 Begin Standard 15W00
 1 Jan 1912 0:00 1:00
 Begin Standard 0W00
 26 Feb 1934 0:00 0:00

which means that Niamey observed a time zone of -0100 (1 hour west of
UTC) between 1912 and 1934.  This is the offset in question.

> The Wikipedia article on Niamey says

Wikipedia itself is not reliable on these issues, unless it cites
sources.

> The PC Atlas uses WAT for UTC-1...
>  
> AstrolDeluxe displays West Central Africa Time (WCAT) = -1...

These are clearly invented abbreviations, just as the TZ abbreviation
is invented.  Microsoft says that the "West Central Africa time zone"
is +0100 (1 hour east of UTC).  See
<http://msdn.microsoft.com/en-us/library/dd796927.aspx>.  All the
other uses of the phrase "West Central Africa time" (that I found with
Google) agree with Microsoft.

> Perhaps a different term (I propose WCAT) would straighten out the ambiguity

"WCAT" would be problematic for reasons described above.  Ambiguity
per se is not a problem in the tz database, as it exists in many other
places (e.g., "EST" in the U.S. versus Australia).  It's not entirely
clear that this Niamey time zone actually existed; it might just be a typo.
Until it's verified, it's not clear that it's worth the effort to
invent a new abbreviation for it.