[Cache from: http://www.y12.doe.gov/sgml/sc34/document/0410.htm, fetched 2003-06-21.]
|TITLE:||Mobile Subset of XML Schema Part 2|
|PROJECT EDITOR:||Martin Bryan|
|STATUS:||For information and review|
|DISTRIBUTION:||SC34 and Liaisons|
We propose to create a compact and reliable subset of W3C XML Schema Part 2 and publish it as an ISO standard. The main target of this subset is mobile devices (such as cellular phones).
Mobile devices are expected to use XML in the near future. Small XML parsers have been developed already. Validators for schema languages are expected to follow, and a prototypical validator for RELAX NG on mobile phones has been developed. Such parsers and validators will hopefully be used for implementing XForms and Web Service on mobile devices.
Part 2 of W3C XML Schema provides a set of datatypes and facets. Although it might not be perfect, it is likely to be widely used by many XML applications including mobile applications. We just cannot believe that an incompatible set of general-purpose datatype (e.g., int) libraries will be accepted by the market.
However, datatypes and facets of W3C XML Schema Part 2 are too complicated for mobile devices. Some specifications such as XForms have already created their own subsets of W3C XML Schema Part 2. However, if different specifications introduce different subsets, incomparability will be significantly spoiled. It would be much nicer if one subset is internationally standardized.
Table 1 gives a list of datatypes of W3C XML Schema Part 2. The exclamation mark indicates those types which we propose to incorporate into the mobile subset.
|3.2 Primitive datatypes|
|3.2.1 string !|
|3.2.2 boolean !|
|3.2.4 float !|
|3.2.5 double !|
|3.2.17 anyURI !|
|3.3 Derived datatypes|
|3.3.1 normalizedString !|
|3.3.2 token !|
|3.3.3 language !|
|3.3.4 NMTOKEN !|
|3.3.5 NMTOKENS !|
|3.3.6 Name !|
|3.3.7 NCName !|
|3.3.8 ID !|
|3.3.9 IDREF !|
|3.3.16 long !|
|3.3.17 int !|
|3.3.18 short !|
|3.3.19 byte !|
|3.3.21 unsignedLong !|
|3.3.22 unsignedInt !|
|3.3.23 unsignedShort !|
|3.3.24 unsignedByte !|
Note: In addition to these datatypes, XForms introduces two groups of datatypes. One group of datatypes contains decimal, dateTime, time, date, gYearMonth, gYear, gMonthDay, gDay, and gMonth. The other group contains integer, nonPositiveInteger, negativeInteger, nonNegativeInteger, and positiveInteger, all of which require infinite precision.
On the other hand, our subset allows normalizedString, token, language, NMTOKEN, NMTOKENS, Name, NCName, ID, IDREF, which are not allowed by XForms. These are not particularly hard to implement. However, XForms people probably did not think that these datatypes are important for XForms.
Table 2 shows facets of W3C XML Schema Part 2. The exclamation mark indicates those facets which we propose to incorporate into the mobile subset.
|4.3 Constraining Facets|
|4.3.1 length !|
|4.3.2 minLength !|
|4.3.3 maxLength !|
|4.3.5 enumeration !|
|4.3.7 maxInclusive !(when datatypes are neither date, dateTime, time, nor duration)|
|4.3.8 maxExclusive !(when datatypes are neither date, dateTime, time, nor duration)|
|4.3.9 minExclusive !(when datatypes are neither date, dateTime, time, nor duration)|
|4.3.10 minInclusive !(when datatypes are neither date, dateTime, time, nor duration)|
We have studied the source code of Jing implementation by James Clark. We believe that if the above restrictions are accepted, an implementation of the remaining datatypes and facets will require less than 20KB as the size of a JAR file.