[
Lists Home |
Date Index |
Thread Index
]
Uche Ogbuji wrote:
> > > I see no reason why the exact same provision cannot be made in the
case of
> > >
> > > <date>2002-08-06</date>
> > >
> > > In the schema, you can say:
> > >
> > > "A date formatted as specified in ISO-8601"
> >
> > yup. but this value _is_ tagged, it's not just a string.
>
> "Just a string" is the heart of this entire argument, and not a given
premise.
> In the end, if it can be expressed in XML, it is "just a string". The
magic
> is in the interpretation, and as long as you have the full tools for
> interpretation, I don't see how gung-ho tagging is any less ambiguous, or
more
> valuable in any other respect, than the use of untagged content. The tags
can
> themselves be the full tools for interpretaion, but in this thread we are
> looking at cases where this is not necessary, because of convention.
Geospace
> and dates are the example.
And my claim is that tagging the data _reduces_ the ambiguity. Now you may
claim that there is no ambiguity in which case you can interpret your
(non-XML) strings. My claim is that if there is ambiguity, tagging reduces
it.
> >
> > Besides the vast majority of Americans write:
> >
> > <date>02-03-2004</date>
> >
> > and so when you have different specs that start mixing the order of
years,
> > months, days it _is_ ambiguous. If you write out the date as is often
seen
> > in legal documents it is the equivalent of the marked up version:
Where is the schema in the above example? I don't see one. I see only
<date>02-03-2004</date>
Now we might use the <date> tag to somehow find a schema but suppose the
naked
02-03-2004
where is the schema? Granted if you have _already ascertained_ that this
value is typed a schema might not be necessary, but is that the case here?
>
> No. It is not the slightest bit ambiguous because the schema
disambiguates
> it. Your example also needs a schema to disambiguate it (as Joe pointed
out,
> you don't state whether it's CE or BCE), so I still don't consider it less
> ambiguous.
Again, what schema?
Suppose the string "10". What does it represent?
a) a title of a movie
b) a decimal age
c) a binary number?
it might be any of the above.
>
> I think Joe made a killer point with his mention that you probably don't
write
> <quantity><hundreds>1</hundreds><tens>4</tens><ones>4</ones></quantity>
I missed the point of this one. All numbers I've seen writted (unlike dates)
are written right to left in increasing significance.
>
> Which I think is the logical conclusion of your position. After all,
> <quantity>144</quantity> could be in octal, hex, or any numerical base
above
> 5, no?
>
"144"
-- could mean either a number or not, might be the name of a truck or
plane, might be a number, could be a code, who knows?
<quantity>144</quantity>
-- the meaning may be constrained by the meaning of the <quantity> element.
The english name of the element i.e. the element type, suggests that the
value is to be interpreted as a number. Usually numbers are written in
decimal but who knows.
<quantity xsi:type="xsd:decimal">144</quanity>
-- no ambiguity in the number, but what the number quantifies might be
anything -- and so we need to look at the context of this.
Which is the logical conclusion.
Jonathan
|
