1. According to official announcement from Korean government, the DST end date in South Korea should be 1955-09-08 without specifying time http://theme.archives.go.kr/next/common/viewEbook.do?singleData=N&archiveEve... 1956-09-29 without specifying time http://theme.archives.go.kr/next/common/viewEbook.do?singleData=N&archiveEve... 1957-09-21 24 o'clock http://theme.archives.go.kr/next/common/viewEbook.do?singleData=N&archiveEve... 1958-09-20 24 o'clock http://theme.archives.go.kr/next/common/viewEbook.do?singleData=N&archiveEve... 1959-09-19 24 o'clock http://theme.archives.go.kr/next/common/viewEbook.do?singleData=N&archiveEve... 1960-09-17 24 o'clock http://theme.archives.go.kr/next/common/viewEbook.do?singleData=N&archiveEve... Contradicting the current tz database record that shows their end date as 00:00 of the next days. The effect should be same in almost all cases, but to be literally reflect the rule in tz database probably an update might be desirable. [Information via Suikawiki] 2. Regarding history of Korean time, there is a Korean wiki site https://namu.wiki/w/대한민국%20표준시 that say when Korea was using GMT+8:30 as standard time, the international aviation/marine/meteorological industry in the country refused to follow and continues to use GMT+9:00 for interoperability. Did the phenomenon actually occured back then and is there any further details about it? 3. Back in 2014 when Korean time history in tz database was discussed, the reliability of The International Atlas by Shanks was also being discussed. I checked the book on amazon.com and there's a review saying that the International Atlas say they've make up some information to fill holes in their research and use them to advance copyright claims to temporarily force tzinfo offline. I am not sure what happened back then in the year of 2011 when the comment was posted, but if that's true then tz database should probably be more actively removing anything related to Shanks from the database as those information are not fact and thus subject to copyright claims and thus cannot be released as public domain according to the license of tz database? Not to mention the accuracy of those data.