LC-27 allowing multiple lexical spaces for floats

Dear Mr. Falk and Mr. de Judicibus:

The W3C XML Schema Working Group has spent the last several months
working through the comments received from the public on the last-call
draft of the XML Schema specification.  We thank you for the comments
you made on our specification during our last-call comment period, and
want to make sure you know that all comments received during the
last-call comment period have been recorded in our last-call issues
list (http://www.w3.org/2000/05/12-xmlschema-lcissues).

Among other issues, you raised (independently) the point registered as
issue LC-27, which suggests that XML Schema provide, or allow schema
authors to provide, different lexical spaces for floating-point
numbers.

We thank you for posing this issue.  In various guises and in
combination with other issues which proved to be related, it has
occupied a great deal of our time during the last-call period.

There is a great deal of sympathy in the WG for providing, in the long
term, systematic support for the multiple different lexical spaces for
simple types, as a method of supporting localized notations for
numbers, non-Gregorian dates, and so on.  On this question your
leanings, the goals of the W3C internationalization WG, and the intent
of the XML Schema WG are all the same.

In the short term, however, it proved rather difficult to achieve
consensus on how best to achieve, or set the stage for later
achieving, those long-term goals.  The input of the W3C
internationalization WG was consistently and strongly against allowing
any value of any simple type to have more than a single lexical form;
their goal as I understand it is to encourage the development of
systems with the ability to localize data formats appropriately for
the particular user they are facing at any given time, and the policy
of 'sender makes right' is a step toward ensuring that such systems
are not too difficult to build.

We are clearly not now in a position to define localized types for
every lexical space now in use to denote common simple types (let
alone those of historical importance which might occur in documents to
be published on the Web).  For a long while, the WG considered a
proposal to define a set of 'abstract' simple types (e.g. an
AbstractFloat type) from which schema authors could derive localized
concrete types (e.g. an AmericanFloat and a EuropeanFloat, using dot
and comma respectively for the decimal point).  In the short term, it
would not be feasible to provide a way for a schema author to define a
mapping between the localized form and the canonical representation of
a given value, so it would be impossible to enforce maximum and
minimum values on such localized types.  This struck some WG members
as taking most of the utility out of the proposal, but some WG members
insisted that it could be useful for an application to know what value
space a given lexical form represented, even without knowing which
value it denoted.  Eventually, a fuller proposal was developed, but
integrating it into the design proved to raise a number of difficult
design questions.  The WG was unwilling to answer these questions
before they had received adequate consideration and the implications
were fully clear, and so in the end the proposal for abstract simple
types, which was intended in part to allow schema authors to define
types for non-gregorian dates (and in the long run to provide a
foundation for standardizing such types) was rejected.  It will not be
part of XML Schema 1.0.

The WG regrets that we were unable to find a way to incorporate a
reliable method for supporting multiple lexical spaces for the same
simple type into XML Schema 1.0; we hope that we will be able to turn
to this question in a later version of XML Schema.

It would be helpful to us to know whether you are satisfied with the
decision taken by the WG on this issue, or wish your dissent from the
WG's decision to be recorded for consideration by the Director of
the W3C.

with best regards,

-C. M. Sperberg-McQueen
  World Wide Web Consortium
  Co-chair, W3C XML Schema WG

Received on Thursday, 5 October 2000 16:58:41 UTC