On Thu, 15 Sep 2011 14:14:02 +0200, [UTF-8?]SaaÅ¡ha Metsärantala wrote
> > to add a little feature.
> If we had an extra week, I would really like considering this interesting
> > y193.9e1 is 1939 with decade precision, the 9 in 1939 is "estimated"
> If we adopt that, maybe we could skip tildes. In this case, implementation
> may (not?) be more difficult, but what about interpretation? If we do not
> skip tildes, we must we VERY clear about the semantic differences between
> the two.
Tilde is an (approximate) estimation on the whole of the date.. It can apply
to day, month, year, time or ...
With scientific notation I am suggesting that only the digits to the right
of the decimal point are estimated. Those to the left are expressed with the
precision of the exponent..
> Furthermore, would 1234.1e0 mean the following "estimation":
> At the begining of the year 1234
I think it is what you wrote.. In the year 1234.. and estimated to be 36
days or so into that year, e.g. end of Jan and maybe the start of Feb.
The expression y2011.08e0 I would take to mean this year (it being 2011) and
estimated towards the end of Jan..
Estimations are just that.. estimatations.. they are good for doing the
match.. 6 months after an event estimated to have been in Jan is quite
different from 6 months after one estimated to have taken place towards the
end of the year... think about reading analog thermometers, volumetric
> Or would it mean someshing else? Or would years with decimal parts be
I was not thinking of them.. but I think they make sense and may indeed be
y2012.5e0 to express the year 2012, estimated the middle of the year..
> If there is no possibility to extend the time dedicated to phase one, I
> would suggest to move this suggestion to phase two (or to phase 1.1 if
> there is such a phase).
> I assume that several of us consider that it would be time to move to a
> phase of "proposed recommendation", and have one more month (or so) for
> implementations and tests before we can recommend the use EDTF. Before
> this week-end, I will not have any possibility to complete the EDTF
> implementation I am currently working on. Of course, I plan to publish my
> source code free of charge for others to use and modify - unless there is
> no more than a weak interest for it - but unfortunately, I won't be able
> to do that before next week (at best).
Edward C. Zimmermann, NONMONOTONIC LAB