[Bug 5192] Terminology: "absent" (editorial)

http://www.w3.org/Bugs/Public/show_bug.cgi?id=5192

           Summary: Terminology: "absent" (editorial)
           Product: XML Schema
           Version: 1.1 only
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Structures: XSD Part 1
        AssignedTo: cmsmcq@w3.org
        ReportedBy: mike@saxonica.com
         QAContact: www-xml-schema-comments@w3.org


The term ·absent· (note the dots) is sometimes used loosely. According to the
definition, the term should only be used as one of the possible values of a
property of a schema component. However, we read for example in 3.2.4:

For an attribute information item to be locally ·valid· with respect to an
attribute declaration all of the following must be true:
1 The declaration is not ·absent· 

This usage suggests that the set of all attribute declarations can be
partitioned into those that are ·absent· and those that are not ·absent·, which
is clearly nonsense. The rule ought somehow to say that it is invoked with a
parameter whose value is either an attribute declaration or the special value
·absent·; or alternatively, that the parameter is optional.

So I looked for places where the rule is invoked. Unfortunately, none of the
places where Attribute Locally Valid is invoked seem to mention the Attribute
Declaration that is passed as the parameter ...

I wrote this thinking I would find other similar examples of misuse of
·absent·, but this seems to be the only one left. There were others in 1.0.

Some other points however:

* The definition of ·absent· suggests it is always a property value. Some
properties however contain a set of values, in which ·absent· can appear
alongside other values. And ·absent· is sometimes used as the value of local
variables or parameters in rules. It might be appropriate to widen the
definition a little.

* When discussing Infoset or PSVI properties, "absent" is usually written
without dots, but in some cases the dots appear (incorrectly?): for example
[expected element declaration] in 3.3.5, "If the [schema normalized value] is
not ·absent· " also in 3.3.5, [document location] in 3.17.5. Perhaps it would
be simpler to use ·absent· (with dots) throughout, and embrace this in the
definition?

3.10.2 has "set consisting ·absent·" for "set consisting of ·absent·"

Michael Kay

Received on Sunday, 14 October 2007 17:02:29 UTC