Africa/Casablanca timezone questions (probably incorrect?)

[GoogleCodeExporter]

Hi,

I have some questions about the Africa/Casablanca timezone. First of all the 
outlook compatible one (http://tzurl.org/zoneinfo-outlook/Africa/Casablanca) 
seems to have a wrong TZOFFSETFROM at the DAYLIGHT timezone part. From my point 
of view the TZOFFSETFROM shoutld be +0000 and not +0100?

However, I have bigger concerns when I look at the none Outlook version 
(http://tzurl.org/zoneinfo/Africa/Casablanca) which I prefer to use at my 
products. 

1) Why does the following timezone start in the year 2026? From my point of 
view it should start at an earlier point in time? Whats your opinion? 
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0100
TZNAME:WEST
DTSTART:20260329T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU

2)The other timezone parts, for my point of view, are also probably not 
completely correct. For example, a standard timezone with an rrule starts in 
the year 2013 (btw. why 2013? shouldn't the same rule also apply at years 
before 2013?):
BEGIN:STANDARD
TZOFFSETFROM:+0100
TZOFFSETTO:+0000
TZNAME:WET
DTSTART:20131027T030000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU

But then there are also other standard timezones which start in 1939 BUT have 
RDATES until 2022 which are quite irregularly defined. Should the standard 
timezone really start on a different month each year? And why are there RDATES 
> 2013 defined when we already have a STANDARD Timezone with a rrule which 
starts in 2013? 
RDATE:20110731T000000
RDATE:20120720T030000
RDATE:20120930T030000
RDATE:20130707T030000
RDATE:20140628T030000
RDATE:20150613T030000
RDATE:20160604T030000
RDATE:20170520T030000
RDATE:20180512T030000
RDATE:20190504T030000
RDATE:20200418T030000

3) Similar questions arise in my mind when I investigate the DAYLIGHT 
timezones. E.g. why does the DAYLIGHT timezone with the rrule start in 2026? (I 
assume that also should start earlier) and also the RDATES show a similar 
unstable occurrence as it does at the example above.

Am I interpreting something wrong here? Please correct me if I do. 

Original issue reported on code.google.com by [email protected] on 26 Feb 2015 at 10:45