Re: XML Schema datatypes

On Thu, 15 Nov 2007, ewallace@cme.nist.gov wrote:
>
> Carsten Lutz wrote:
>
>> On Tue, 13 Nov 2007, Jeremy Carroll wrote:
>>>
>>> 7) rounding errors behave very differently from in traditional numeric
>>> applications - hence a solved problem (rounding) becomes an unsolved problem
>>>
>>> (I will also add another area of concern which is a mismatch between the
>>> real numbers and the XSD datatypes)
>>>
>>> ===
>>
>> This is an interesting subject. I don't think it is really about
>> "rounding", though, and neither is it really about n-ary datatypes.  I
>> think what your example shows is that using bounded or fixed-precision
>> datatypes such as float or decimal in the semantics of an ontology
>> language is a bad idea. This means that already the unary datatypes in
>> OWL 1.0 are in some sense broken, and the XML Schema datatypes are
>> probably *not* the right thing to use in OWL. Sorry if this sounds like
>> blasphemy.
>
> A minor nit:
>
> The built-in XML Schema datatype Decimal is neither bounded nor fixed
> precision (as defined).  The type is basically the Reals subsetted to those
> values that can be expressed with finite length decimal numerals.

I have to admit that I wasn't aware of that. As far as I can see, I
have no objections against XSDs integer. It qualifies as a
"mathematical datatype" in the sense of my previous mails.  Also XSD
strings seem fine, though I wasn't able to check all details yet.

So I have to update my view on XSD a bit: it is certainly tailored
towards machine datatypes, but there are also datatypes in it that go
along well with OWL. So I modify my "XSD is probably not the right
thing for us" statement to "XSD is not enough" and "many XSD datatypes
may yield unintuitive result, and this is an issue that we should in
some way address".

>  So
> in the terminology that Jeremy used in an earlier email, it is sparse
> in comparison to Real or even Rational numbers.  I believe that this was
> a type that was meant to support a wider range of uses, where the
> precision and range would be bounded according to the needs of the
> application, not the representation.

That's exactly the point I wanted to make.

>  It's closer to what we need then
> standard computational types like Float, but still not good enough.
>
> I agree with Carsten though that the issue here isn't really rounding.
> There were three types of concerns in the Turner - Carroll paper
> that were mixed:
> 1) computational differences/limitations in types,
> 2) representational differences/limitations in types and
> 3) naive expectations of users.

Again, this is also what I feel.

> I would like to see us explore defining a Rational datatype and
> corresponding computation requirements for reasoners to address limitation

I agree. I would very much like to see in OWL 1.1 datatype for the
rationals. Regarding the reals, there is of course the problem of
representing constants, but I don't think that is prohibitive and we
may consider adding the reals as well.

greetings,
 		Carsten

--
*      Carsten Lutz, Institut f"ur Theoretische Informatik, TU Dresden       *
*     Office phone:++49 351 46339171   mailto:lutz@tcs.inf.tu-dresden.de     *

Received on Friday, 16 November 2007 07:57:31 UTC