And putting CET for a place observing DST is misleading, at least.
Europe/Berlin would be correct, with CET/CEST combining Europe/Berlin, Europe/Brussels, etc…
… or deleting them, as so far nobody could provide any verifiable source about those identifiers. Still waiting for that.
Regardless, please get rid of the separate relation of Europe/Berlin, as it’s a duplicate of an administrative boundary, and therefore should not exists according to the wiki. Whole Germany is within one zone only.
You might want to reach out to German Federal Government and tell them ;) It was not me wrote the law about CET being the official timezone in Germany.
From your wiki-page changes:
This database uses geographical identifiers with a fixed notation in the form of
Continent/Capital— for example,America/New_York,Europe/Amsterdam, andAsia/Pyongyang.
Afaik, New York is not a capital. Evan the capital of the state New York is Albany. You might want to pick another example. As well you might want to explain how to get to the City name, if its not a capital, like in America/New_York
What exactly do you want? Some people used timezone=* to record the IANA identifiers that correspond to those regions. There is absolutely no doubt about which area corresponds to Europe/Berlin (continental Germany, that is). Demanding a verifiable source is just silly for something as widely used as this.
I agree with you that those are not actual timezones, but rather convenient pointers which in tz database always point to the current rule set.
They are used in OpenStreetMap, so you will have to deal with them. Or if you want to record actual standard timezones like CET/CEST, ignore them completely and use different tags. Starting a mapping war about the definition of the key timezone is not really useful.
Afaik, New York is not a capital.
Edited.
You just deleted it I think. Agreed, that is a pointless relation.
Nope, I just deleted the Europe/Old Berlin and also added Europe/Berlin to Germany and shifted CET into standard_time as suggested by you. Relation: Europe/Berlin Timezone (15863720) | OpenStreetMap is still existing and spanning Germany and whole Northern Europe. If that can be deleted, I can do so.
In Europe it might be a bit pointless, if I should add a tag <continent>/<capital> to a country. But if that’s pointless, why do we need to tag it? Software can derive in which continent Germany is and whats the capital of it. Don’t you agree?
But take a look to the US, where it’s totally cluttered. Where is America/New York ending? Or take a look to Asia. Why it is Asia/Shanghai? Capital is Beijing, largest City in China is Chongqing.
Because those IANA identifiers are broadly used. Compare it to ISO3166-1:alpha2. Used everywhere, so it makes sense to tag the areas they correspond to.
If you don’t want to deal with that complexity, don’t. You can probably find out in the tz database files which are annotated with a bunch of comments.
That is similar to the Europe/Brussels I deleted. Again, a cluster of countries which only exist in tz database for deduplication reasons. I agree that no matter how you define Europe/Berlin, it does not include Sweden at the moment — Germany just isn’t as expansionist as it was these days.
That one makes no sense in the context of OpenStreetMap. You could contact the initial mapper as a courtesy, and delete it if you get no reply (I didn’t receive a reply on my comment for the BeNeLux spanning Europe/Brussels either, despite the mapper being active according to their changeset history).
You mean half of the time, Germans don’t follow Germen law? 
1 Like
See…
As of today, whole Germany is following one time zone.
Are you intending to change the relation twice a year???
I thought it was just silly to have a source, as it’s something obvious.
Now it seems super complex. If someone is adding data to OSM, he should be able too provide a valid source.
If the law is changing, there will be updates needed. Current law is from 1985, so I don’t expect many changes.
You said CET, but in summer Germany is using CEST, not CET.
Either get rid of those identifiers. Otherwise the absolute bare minimum would be that those people, who dumped that data properly document the tag within the osm-wiki, in a way that other mappers can validate the correctness of that data.
Part of that documentation should be how those areas can be identified. As you stated, some of those areas are trivial, like Europe/Berlin. But I pointed out several values, which are not.
I don’t get your point. The official timezone is CET.
That’s what I initially tagged. Then @JeroenHoek suggested an alternate tag, since timezone is occupied now by some identifier and as well suggested how to tag the fact, that there is a DST. I changed to those tags…
Nach § 4 Abs. 2 EinhZeitG aufgrund einer Verordnung des Bundesministeriums für Wirtschaft und Energie in einem Zeitraum zwischen Ende März und Ende Oktober die gesetzliche Zeit die Mitteleuropäische Sommerzeit (MESZ), definiert als Koordinierte Weltzeit (UTC) plus 2 Stunden.
Maybe read the actual law, not wikipedia?
§ 4 Gesetzliche Zeit
(1) Die gesetzliche Zeit ist die mitteleuropäische Zeit. Diese ist bestimmt durch die koordinierte Weltzeit unter Hinzufügung einer Stunde.
(2) Für den Zeitraum ihrer Einführung ist die mitteleuropäische Sommerzeit die gesetzliche Zeit. Die mitteleuropäische Sommerzeit ist bestimmt durch die koordinierte Weltzeit unter Hinzufügung zweier Stunden.
Before @JeroenHoek suggested to tag any DST, there was only one tag for timezone, which in general is CET, CEST is a exception for a certain period of time.
"mitteleuropäische Sommerzeit ", eben!
For better or worse, the boundary=timezone and timezone=* pages have specified IANA time zone identifiers since 2017.
An IANA time zone name consists of the file name (e.g., Pacific) followed by one or more place names. Some names have three parts, like America/Indiana/Tell_City. The last part isn’t necessarily a capital or even a city. (America/New_York does refer to New York City, also known as New York.) When a city name is used, it’s generally the most populous city in the zone, even if much of the zone is in a different country than that city. This is just a naming convention. The name Europe/Berlín doesn’t imply that a given location within the zone answers to Berlin in any way. Even when a zone exactly corresponds to one country, the database avoids the country name, in case the country changes its name.
The tz theory document discusses the naming convention in more detail. The maintainer is authorized to choose time zone names by an IETF standard.
None of these details are particularly relevant to this discussion. As long as we include an external identifier in OSM, we should aim to indicate the correct identifier, consistent with the external source. Otherwise the external identifier would be worse than useless – it would be misleading. For all we care, the tz maintainer could choose an atypical name for a new zone and I guess that’s what we’d go with, if anything at all.
If by verifiability you mean that someone can verify the truth, then anyone can consult the tz database and associated discussions on the tz mailing list. There’s no secret conspiracy to defend against.
On the other hand, if you mean observability, then you must surely think that actual time observance is more relevant to OSM than the text of a law. If so, then you probably favor mapping the IANA time zones, which aim to reflect actual time observance rather than the often ignored text of a law. After all, if some region officially observes one standard time but actually observes a different standard time in practice, then the standard time’s name is misleading or doesn’t tell the whole story.
4 Likes