Dave Pawson:
> Outsider view.
> Not a solution. Just more of an interop problem.
> Use iso8601 inclusive of TZ.
We already use ISO 8601, which supports indicating timezones.
Timezone information is _optional_ in ISO 8601, though,
so _not_ indicating the timezone is _still_ compatible with ISO 8601.
There is no timezone data in the built-in "Date" type of spreadsheets.
Writing out a timezone marker (such as "Z" for UTC) would be WRONG,
because there is NO such information in a spreadsheet.
I think it'd be great to extend spreadsheets to provide proper
support for timezones, but that turns out to be:
1. Nontrivial.
2. Implemented by NO spreadsheets
I think it'd be inappropriate to do major surgery on a basic
type when there's no experience with such changes. Let's standardize
on where there _IS_ agreement first.
I can imagine ways to add timezone information.
Perhaps there could be an additional value attached to the table
or sheet that would indicate the epoch and "default display"
timezones [+/-]xx:yy. Then you could format displays to
use the "epoch timezone", "default display timezone", or
"UTC". Another approach, which isn't
inconsistent with this, would be to add functions to do timezone
conversions into strings (so display can be controlled).
Others can probably imagine several more.
But I think it would _wrong_ to standardize what are, essentially,
unproven experiments.
--- David A. Wheeler
