Print

Print


Continuation off topic... Thats correct with the word century and
swedish. And yes its a common missunderstanding but its really a common
one here.

>>> John Hostage <[log in to unmask]> 2015-05-19 15:32 >>>

This is a bit off-topic, but I think this is a misunderstanding of the
meaning of century.  In English we can say “the 20th century” or “the
1900s”.  They mean the same thing.  As I understand it, Swedish doesn’t
have a word for century in this sense.  You can only say “1900-talet”. 
However you say it, it comes out as “19” in ISO 8601.
 

------------------------------------------
John Hostage
Senior Continuing Resources Cataloger
Harvard Library--Information and Technical Services
Langdell Hall 194
Harvard Law School Library
Cambridge, MA 02138
[log in to unmask]
+(1)(617) 495-3974 (voice)
+(1)(617) 496-4409 (fax)
ISNI 0000 0000 4028 0917

 

From: Discussion of the Developing Date/Time Standards
[mailto:[log in to unmask]] On Behalf Of Karin Bredenberg
Sent: Tuesday, May 19, 2015 01:11
To: [log in to unmask]
Subject: Sv: Re: [DATETIME] Long Years vs. Expanded Representations;
forms with century vs. year precision

 

Just need to add, being from Sweden when we say 19th century (in
Swedish: 1900-talet) we mean 1900-1999.

 

Best,

Karin

>>> John Hostage <[log in to unmask]> 2015-05-18 18:12 >>>

According to ISO 8601, sec. 4.1.2.3 (Representations with reduced
accuracy), centuries are written with two digits.  So 20th century = 19
 
Probably not very satisfactory in some contexts.
 
------------------------------------------
John Hostage
Senior Continuing Resources Cataloger
Harvard Library--Information and Technical Services
Langdell Hall 194
Harvard Law School Library
Cambridge, MA 02138
[log in to unmask]
+(1)(617) 495-3974 (voice)
+(1)(617) 496-4409 (fax)
ISNI 0000 0000 4028 0917
 
From: Discussion of the Developing Date/Time Standards
[mailto:[log in to unmask]] On Behalf Of Nathan Harrenstein
Sent: Friday, May 15, 2015 01:33
To: [log in to unmask]
Subject: Re: Long Years vs. Expanded Representations; forms with
century vs. year precision
 

Yes, you are both right. It was an error on my part; nevertheless, some
notation could be used to differentiate centuries and years. Centuries
can be written [1900..1999] using the current EDTF extensions, or a
notation could be invented. Then:

 

20th century = [1900..1999]

Year 19000  = +19000

Date with the year 19000 = +19000-12-25

 

I don't know if there would still be a reason for positive expanded
representations to be prefixed with a "+", but maybe they could be
omitted.

 

Nathan

 

On Thu, May 14, 2015 at 5:48 PM, Byrd, Donald A. <[log in to unmask]>
wrote:
Since Edward Zimmerman hasn't spoken up, I'll jump in and say I also
feel strongly that the semantics of  "(NN+1) century" and "NNxx", where
the N's are digits, are not identical. The former has a precision of a
century, the latter a precision of a year. For example, the American
Revolution occurred in the 18th century; since it lasted much longer
than a year,  saying it occurred in 17xx is simply wrong. On the other
hand, for when the composer Perotin was born, you could equally well say
either 11xx or the 12th century.

--DAB


On May 11, 2015, at 4:58 PM, "Denenberg, Ray" <[log in to unmask]> wrote:

> Yes, makes complete sense now, and thanks.
>
> The possibility of using masked precision in the manner you suggest
was not discussed (as I recall), but there is a reason why it would most
likely have been rejected.  Basically, the semantics of ‘19xx’ would not
quite be the same as the semantics of “20th century”. The reason has to
do with the concept of “precision”.
>
> We do have a resident expert on “precision” on this listserv, Edward
Zimmerman, I hope he’s listening and if so, certainly he will
elaborate.
>
> Ed was(is)  quite passionate about the concept of precision, and
while not all of us shared his passion, he was a significant contributor
to the spec and we deferred to his expertise in this matter.
>
> The argument would go  something like this.   ‘19xx’ represents a
year: an unspecified year within the 20th century.  The semantics of
“20th century” are not quite so specific.
>
> Hope that helps, and hopefully Ed will weigh in.
>
> Ray
>
> From: Discussion of the Developing Date/Time Standards
[mailto:[log in to unmask]] On Behalf Of Nathan Harrenstein
> Sent: Saturday, May 09, 2015 12:15 AM
> To: [log in to unmask]
> Subject: Re: [DATETIME] Long Years vs. Expanded Representations
>
> In ISO 8601:2004(E), dates can be expressed in either "complete
representations" or "representations with reduced accuracy". Dates of
the form "±YYYYY-MM-DD" or "±YYYYYMMDD" are called complete
representations. The day, the month, or the latter half of the year may
be omitted from a complete representation, in which case the date is
referred to as a representation with reduced accuracy. This includes a
specific month "±YYYYY-MM", a specific year "±YYYYY" or a specific
century "±YYY".
>
> This feature requires the agreement on the number of added digits
between interchange parties because otherwise it would not be clear if a
date such as "+00019" referred to the 20th century or the year 19. If it
was known that the time element was expanded by three digits, then
"+00019" would mean the 20th century because adding three digits to five
means that there must have been two digits to begin with, and a year
must have at least four digits. However, if the time element is known to
have been expanded by one digit, then "+00019" would mean the year 19
because there must have been four digits to begin with. Without knowing
how many digits were added, the expression "+00019" is ambiguous and
could represent either the 20th century or the year 19.
>
> The EDTF "long year" feature with the "y" prefix does not support
representations with reduced accuracy insomuch as months, days, and
centuries are unsupported (e.g. y00019 would never mean the 20th
century). Because long years do not support these forms of reduced
accuracy, there is no ambiguity and therefore no need for prior
agreement on the number of digits between interchange parties.
>
> Does this accurately summarize the rationale for adding the long year
feature? Would it not be simpler to alter expanded representations such
that the centuries are notated using masked precision and only the
extended format of expanded representations are supported (i.e.
±YYYYY-MM-DD is supported but not ±YYYYYMMDD)? For example, "+019xx"
would refer to the 20th century, "+00019" would refer to the year 19,
and a full date would be written "+00019-12-25". Then the feature that
already exists is improved instead of working in another similar
feature.
>
> Hopefully some of this makes sense,
> Nathan
>
> On Fri, May 8, 2015 at 12:19 PM, Denenberg, Ray <[log in to unmask]>
wrote:
> I’m sorry but I don’t understand the first question.
>
> On the second, the need for prior agreement certainly makes a spec
less interoperable (if not nonstandard) which is one reason we came up
with an alternative approach.
>
> Ray
>
> From: Discussion of the Developing Date/Time Standards
[mailto:[log in to unmask]] On Behalf Of Nathan Harrenstein
> Sent: Thursday, May 07, 2015 2:49 PM
>
> To: [log in to unmask]
> Subject: Re: [DATETIME] Long Years vs. Expanded Representations
>
> So if I understand correctly, long years do not support any form of
reduced accuracy to avert any ambiguities between centuries and years
that would necessitate an agreement on the number of digits between
parties?
>
> Wouldn't the need for prior agreement between parties make expanded
representations, in a sense, nonstandard?
>
> Nathan
>
> On Thu, May 7, 2015 at 7:15 AM, Denenberg, Ray <[log in to unmask]> wrote:
> 'otherwise a string like "+1919" might mean the year 1919 or the
century beginning in 191901. I think.'
>
>
>
> Small point, but I would say " a string like 191901 might mean the
year 191901 or it might mean January, 1919".
>
>
> Ray
>
>
>
>
> > -----Original Message-----
>
> > From: Discussion of the Developing Date/Time Standards
>
> > [mailto:[log in to unmask]] On Behalf Of Byrd, Donald A.
>
> > Sent: Thursday, May 07, 2015 9:38 AM
>
> > To: [log in to unmask]
>
> > Subject: Re: [DATETIME] Long Years vs. Expanded Representations
>
> >
>
> > I Even aside from the "exponential" notation, these features aren't
the same.
>
> > ISO 8601 says years with more than four digits must be agreed on by
the
>
> > parties, and it strongly implies that an exact _number_ of
additional digits
>
> > must be agreed on. (Which makes sense in that otherwise a string
like
>
> > "+1919" might mean the year 1919 or the century beginning in
191901. I think.)
>
> > In EDTF, years with more than four digits can always be used, and
there's no
>
> > need for (or point to!) agreement between the parties. So EDTF's
>
> > representation is more standardized and well-defined -- and much
better,
>
> > IMO.
>
> >
>
> > --Don
>
> >
>
> >
>
> > On May 7, 2015, at 1:14 AM, Nathan Harrenstein
>
> > <[log in to unmask]> wrote:
>
> >
>
> > > I was reading the ISO 8601:2004(E) standard and noticed that
there is
>
> > already a feature for representing long years, called "expanded
>
> > representations" (§4.1.2.4). Reading through the EDTF listserv
archives, I got
>
> > the impression that this feature was not known at the time long
years were
>
> > ideated.
>
> > >
>
> > > "By mutual agreement of the partners in information interchange,
it is
>
> > > permitted to expand the component identifying the calendar year,
which is
>
> > otherwise limited to four digits. This enables reference to dates
and times in
>
> > calendar years outside the range supported by complete
representations, i.e.
>
> > before the start of the year [0000] or after the end of the year
[9999]" (§3.5).
>
> > >
>
> > > To specify an "expanded representation", the year is prefixed by
a "+" or "-
>
> > ". It doesn't cover the scientific notation features, however.
>
> > >
>
> > > Am I missing something or are these features the same?
>
> > >
>
> > > Nathan
>
> >
>
>

---
Donald Byrd
Woodrow Wilson Indiana Teaching Fellow
Adjunct Associate Professor of Informatics
Visiting Scientist, Research Technologies
Indiana University Bloomington