[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: Tags and Types (was Re: Re: maps)

• To: xml-dev@l...
• Subject: Re: Tags and Types (was Re: Re: maps)
• From: Joe English <jenglish@f...>
• Date: Tue, 06 Aug 2002 15:33:04 -0700
• In-reply-to: <012f01c23d8e$1f1a1a90$0201a8c0@n...>
• References: <012f01c23d8e$1f1a1a90$0201a8c0@n...> <00c801c23d80$d1c21160$0201a8c0@n...> <E17c9p6-0007Ts-00@m...> <200208062107.g76L7N332434@d...>

Jonathan Borden wrote:
> Joe English wrote:
> >
> > In the first example, did you mean 1 CE or 2001 CE?
> > If the former, under which calendar should the day
> > and month be interpreted?
>
> I've not specified, but that isn't needed. The former example is tagged, and
> so the "01" may easily be interpreted with respect to the <year> element,
> and its associated semantics [...]
>
> > Sorry, I really don't see any decrease in ambiguity
> > in the first form.
> Ambiguity is _decreased_ not eliminated because aside from which year we are
> talking about we know which lexical value is intended to represent _some
> year_ as opposed to _some day_.
>
> If it didn't reduce ambiguity _some_ then there would hardly be a need for
> XML, or SGML for that matter.


I guess it's that most of the time I prefer to treat dates as
atomic values; for that purpose <date>2001-02-03</date> is just
as good as <date><year>2001</year><month>2</month><day>3</day></date>,
just as <quantity>144</quantity> is preferable to

<quantity><hundreds>1</hundreds><tens>4</tens><ones>4</ones></quantity>.

I also use <author>Joe English</author> more often
than <author><givenName>Joe</givenName><surname>English</surname></author>,
for much the same reason.

But this all depends on the application.  When the anticipated
processing involves accessing the subcomponents of a compound 
data type, then it's useful to explicitly mark them up;
but if it doesn't then there's little need to do so.

For example, in a bibliographic database where I might need to
format authors' names differently depending on the bibliography
style, explicit markup for first and last names would be
appropriate.  OTOH, in a bug-tracking database treating bug
submitters' names as atomic strings is just fine.  Similar
considerations are involved in the choice between atomic dates and
separate markup for year, month and day, whether or not it's
necessary to specify the calendrical system, et cetera.

And under NO CIRCUMSTANCES would I ever allow "01" to be used
as an abbreviation for "2001 CE" in any markup vocabulary I design :-)


--Joe English

  jenglish@f...

• Follow-Ups:
  • Re: Tags and Types (was Re: Re: maps)
    • From: "Jonathan Borden" <jborden@a...>

• References:
  • Re: Tags and Types (was Re: Re: maps)
    • From: "Jonathan Borden" <jborden@a...>
  • Re: Tags and Types (was Re: Re: maps)
    • From: "Jonathan Borden" <jborden@a...>
  • Re: Tags and Types (was Re: Re: maps)
    • From: Uche Ogbuji <uche.ogbuji@f...>
  • Re: Tags and Types (was Re: Re: maps)
    • From: Joe English <jenglish@f...>

• Prev by Date: Re: Deterministic content models???
• Next by Date: XSL Transform special character into <break>
• Previous by thread: Re: Tags and Types (was Re: Re: maps)
• Next by thread: Re: Tags and Types (was Re: Re: maps)
• Index(es):
  • Date
  • Thread