Stephen Colebourne wrote:
As I and others have indicated before, this change should not have happened. The likelihood of it breaking something was always very high, and the alleged reality is that it expresses is a fact that is irrelevant to most users (and is disputed anyway). This was change for changes sake.
I disagree. This was change in order to be more accurate: if we're offering the isdst flag then it ought to reflect what was actually happening in each jurisdiction. By its nature this flag is less objectively defined than the other outputs of the database, but in this case we have clear evidence that it should be a particular way round. There was indeed a significant chance of the negative SAVE value breaking alternate zic implementations, and now that that's turned out to be the case that's good cause to postpone this correction, or at least the expressing of it in this natural manner. The tzdb does have unusually high stability requirements. But that's not a reason to keep the database forever inaccurate. -zefram