We can, I think, argue that their semantics for accuracy assumed zero bias and
subsummed BOTH precision (readability) and accuracy.
ISO 5725 defines accuracy to mean the closenest to the true value and
precision as the closeness of agreement among a set of results. Should one
take every expression of date as known and true they, of course, align.
We have now, however, extended things in EDTF to contain concepts such as
uncertainty, belief, information, knowledge etc. and thus must explicitly
seperate the two concepts. As I suggested I'm, in fact, not sure if we should
even continue to speak of "accuracy" in the context of ISO 8601 as we have no
ground truth, no universal consensus of what a true value can be. What remains
is our concept of precision (readability) and our belief. I understand that
following that path might be "too philosophical" for some...
But.. I think wecan also argue that we need to pull accuracy and precision
apart alone due to our use of unknowns which were not art of 8601.
Look at date expressions like 196u-12-12. To speak of that date as accurate to
day is completely foolish. We don't even know the year. It's accuracy is
within a decade. The "true" date might end up being 1962-12-12 but it could
also end up as 1969-12-12. We know it was 12 Dec. We know that it was in the
1960s. While an expression such as 1969-12-12 could be in this light
considered as both precise to date and accurate to date all those other
expressions where we have unknowns or express uncertainty etc. don't and can't
show this kind of conceptual aligment...
On Mon, 8 Feb 2016 10:57:09 -0500, Denenberg, Ray wrote
> [UTF-8?]“I find it troubling that an ISO committee is confusing terms
taught in primary school [UTF-8?]science.”
>
> Just to be clear, it [UTF-8?]isn’t the ISO committee [UTF-8?]that’s
confusing these concepts [UTF-8?]– [UTF-8?]it’s ISO in general. The
committee draft has been reviewed by ISO editors (outside the committee) who
have suggested that we change terminology (in the new material) to conform to
existing ISO terminology, which uses [UTF-8?]“accuracy” rather than [UTF-
8?]“precision”. For example, see the excerpt below from ISO 8601-2004
[UTF-8?]“Complete [UTF-8?]Representation” vs. [UTF-8?]“Representation
with reduced [UTF-8?]accuracy”. [UTF-8?]“Accuracy” here should instead
be [UTF-8?]“precision”, but it has been [UTF-8?]“accuracy” since the
first version of 8601 (many years ago) and nobody has noticed (or perhaps
nobody cared enough, since the misuse of the terminology really has no effect
on interoperability). Our committee is not tasked with nor does it have the
authority to fix this sort of thing, however, we are arguing that that does
not mean we have to carry this mistake over to new material.
>
> I just wanted to make it clear that the members of the current committee do
understand the difference between these concepts.
>
> Ray
>
>
______________________________________________________________________________
___________
> 4.1.2.2 Complete representations
> When the application identifies the need for a complete representation of a
calendar date, it shall be one of
> the numeric expressions as follows, where [YYYY] represents a calendar year,
[MM] the ordinal number of a
> calendar month within the calendar year, and [DD] the ordinal number of a
calendar day within the calendar
> month.
> Basic format: YYYYMMDD Example: 19850412
> Extended format: YYYY-MM-DD Example: 1985-04-12
> 4.1.2.3 Representations with reduced accuracy
> If in a given application it is sufficient to express a calendar date with
less accuracy than a complete
> representation as specified in 4.1.2.2, either two, four or six digits may
be omitted, the omission starting from
> the extreme right-hand side. The resulting representation will then indicate
a month, a year or a century, as
> set out below. When only [DD] is omitted, a separator shall be inserted
between [YYYY] and [MM], but
> separators shall not be used in the other representations with reduced
accuracy.
> a) A specific month
> Basic format: YYYY-MM Example: 1985-04
> Extended format: not applicable
> b) A specific year
> Basic format: YYYY Example: 1985
> Extended format: not applicable
> c) A specific century
> Basic format: YY Example: 19
> Extended format: not applicable
--
Edward C. Zimmermann, NONMONOTONIC LAB
|