[tz] Confusion about links in backward and backzone files

[Jon Skeet]

I've just been trying to "compile" the 2014g data into my Noda Time data
format, and I've run against a couple of problems due to trying to include
both the "backward" and "backzone" files.

I'd like to just clarify the intention here, from the perspective of direct
consumers of the files providing a service to others. It looks (from a
brief look) like everything that's in backzone is also in backward, but the
reverse is untrue - so it's not like we should *always* use backwards, but
optionally *override* some previous definitions using the data in
backzone... so things that would otherwise be defined as aliases may gain
full zone rules, and aliases themselves may change (e.g.
Antarctica/South_Pole). Is that a fair representation?

I suspect that for the moment I just won't include backzone as a data
source, with the result of the same set of zone IDs as in 2014f, but more
of them being aliases. For most Noda Time users that's probably fine, but I
may well want to allow users to use backzone data in the future, so I
figure it's worth understanding as early as possible...

Jon
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mm.icann.org/pipermail/tz/attachments/20140902/f865831b/attachment.html>