Oops! in E2-44

(The following substantive problems were brought to my attention by Sandy
Gao in
http://lists.w3.org/Archives/Member/w3c-xml-schema-ig/2003Feb/0246.html,
a posting on the members-only IG list.)

When (last July!) we approved the text for E2-44, we failed to notice
that a corresponding change was needed to 4.3.11.3 totalDigits Validation
Rules, which still reads:

>Validation Rule: totalDigits Valid
>A value in a ·value space· is facet-valid with respect to ·totalDigits· if:
>
>1 the number of decimal digits in the value is less than or equal to {value};

Editorial:  I assume the semicolon at the end should be a period.  Should
the rule be cast as a list when there is only one list item?

Substantive:  The list item still refers to "the number of decimal digits
in the value"; the primary reason for the change was to make clear that
we are talking about *decimal numbers*, not numerals--numbers don't have
digits in them.

The formula relating numbers in the value space to the totalDigits value
is now given in the definition of totalDigits, in 4.2.11.  Surely this
formula need not be repeated; I suggest replacing "1 the number of decimal
digits in the value is less than or equal to {value};" with "1 the value
satisfies the constraint in the definition of totalDigits.", (where the
final 'totalDigits' should reference the definition in 4.2.11).

A parallel change should be made in 1.2.12.3.

Substantive (perhaps):  In 4.2.11 a new paragraph was inserted:

>The term totalDigits is chosen to reflect the fact that it restricts
>the ·value space· to those values that can be represented lexically
>using at most totalDigits digits. Note that it does not restrict the
>·lexical space· directly; a non-·canonical lexical representation·
>that adds additional leading zero digits or trailing fractional zero
>digits is still permitted.

This wording suggests that the representations referred to in "that can
be represented lexically..." are canonical.  This is not so; e.g., 0.1
qualifies for totalDigits = 1 (0.1 does satisfy the formula added in
the erratum).  ('.1' is such a representation; the canonical one is not.)

I recommend removing "non-·canonical" in its only occurrence in this
paragraph; the canonical representation '0.1' does have one "additional
leading zero digit", and is therefore permitted.
-- 
Dave Peterson
SGMLWorks!

davep@acm.org

Received on Thursday, 27 February 2003 16:06:23 UTC