[tz] Turkey delays winter time
eggert at cs.ucla.edu
Mon Sep 21 21:48:18 UTC 2015
On 09/21/2015 01:41 PM, Deborah Goldsmith wrote:
> when you’re pushing bits to hundreds of millions of devices you need
> to do thorough testing. There are also release cycles, as users don’t
> like to be bombarded with requests to update their devices too frequently
Red Hat and Ubuntu have tens of millions of users. This difference in
scale does not explain why Apple can take over six months to propagate a
small change to time zone data, whereas Ubuntu can do it in a few
hours. More plausibly, the difference comes from Apple bundling time
zone data into its operating systems and requiring a full test cycle and
OS upgrade in order to install fixes, whereas Red Hat, Ubuntu, etc.
treat time zone data separately and routinely issue minor updates for
just a few small files, which can be tested separately.
The incremental approach is better for end users and for developers and
for testers, and one way or another Apple needs to change its software
process to support it. Any such change is needed regardless of what
tzdata does, because governments typically don't notify us a year in
advance. Of course a development-organization battleship cannot turn on
a dime, but that's OK; there's no rush, as in the meantime you can
cherry-pick changes from the experimental version on Github as needed.
> If there is impact on IANA from releasing more frequent updates
There is some impact on IANA, though not much: they are mostly just
hosting the data. There is more impact on me, and on development
organizations downstream from tzdata. Every time I do a release, I
check over the distribution; some of this work is automated and some is
not. My work is volunteer and my time is limited. The situation is
similar for many other downstream developers. Most do not have the
resources of Ubuntu or Red Hat, much less Apple or Google.
More information about the tz