Re: Implemenation flexibility for representations of unbounded properties

Expanding just a little bit on Dan's answer, one the reason that the 
workgroup decided that support of arbitrary size bounds was a reasonable 
requirement is that most of the environments in which we anticipate 
schemas will be implemented (e.g. Java) have readily available libraries 
for dealing with such numbers.  Furthermore,  a number of languages over 
the years have successfully used very large or unbounded numbers in 
situations that appear to be more performance critical (e.g all arithmetic 
in the REXX language) than our use in schemas. Our design is in part 
justified by their experiences, and it also maximizes our ability to 
interact with such languages and systems that themselves use large 
numbers.

So, as Dan points out, the workgroup would like to receive feedback from 
actual implementation experience to determine what the overhead and 
complexity proves to be.

------------------------------------------------------------------------
Noah Mendelsohn                                    Voice: 1-617-693-4036
Lotus Development Corp.                            Fax: 1-617-693-8676
One Rogers Street
Cambridge, MA 02142
------------------------------------------------------------------------







Dan Connolly <connolly@w3.org>
Sent by: xmlschema-dev-request@w3.org
05/09/00 04:34 PM

 
        To:     Peter Canning <canning@vitria.com>
        cc:     xmlschema-dev@w3.org, (bcc: Noah Mendelsohn/CAM/Lotus)
        Subject:        Re: Implemenation flexibility for representations of unbounded properties

Peter Canning wrote:
> [...]
> Are implementations required to
> support arbitrarily large values for the "min occurs" property?

Yes, per the current draft as written.

>  If not
> what are they required to support?
> 
> The similar situation obviously occurs for string valued properties, and
> probably for others
> 
> It would be helpful if the specification gave some guidance in the
> requirements for conformance is these kinds of areas.

Actually, we're hoping the implementors will inform the specification,
at this point:


                 NOTE: The use of arbitrary precision
              decimal numbers, including all datatypes
              derived from decimal (e.g., integer) in this
              design impacts the implementation of
              schema processors in a number of places:
              checking maxLength constraints on strings,
              for example. It may impact interchange
              between XML schemas and programming
              languages, databases, etc. 

              Our design discussions did not reveal
              convincing evidence of undue burden
              because of arbitrary precision decimal
              numbers in this design, but we welcome
              further input from implementors. 

                 -- http://www.w3.org/TR/2000/WD-xmlschema-2-20000407/#decimal


-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/

Received on Tuesday, 9 May 2000 18:29:24 UTC