Tony Graham is a staff engineer with Sun Microsystems Ireland. He is a current member of the W3C XSL Working Group, and his involvement with XML goes back to the W3C "SGML WG" that advised the "SGML ERB" (which became the XML SIG and XML ERB, respectively). Tony is also the author of "Unicode: A Primer" (IDG Books, 2000) and of articles, tutorials, and conference presentations on Unicode, XML, and XSL. An Australian, he previously worked in Japan for four years producing SGML documents in English, Japanese, Chinese (Simplified and Traditional), and Korean, and then in the U.S. for five years as an SGML and XML consultant.
Sun's expectation of the output of the workshop is a focused list of objectives for the I18N WG that will provide the basis of the rechartering of the WG when its current charter expires in March 2002.
There is a need for guidelines to aid the stylesheet authors in the development of localised/localizable XSL and XSLT stylesheets.
There is a related need for guidelines for tool developers on how to incorporate locale-sensitive collation and numbering.
XSLT, for example, supports specification of language-specific sort order within xsl:apply-templates and xsl:for-each elements. The current XSLT 2.0 Working Draft, however, notes "It is possible for two conforming implementations not to sort exactly the same." The current WD also weakly notes "It is recommended that implementors consult [UNICODE TR10][5] for information on internationalized sorting".
Similarly, XSLT supports language-specific number-to-string conversion with the xsl:number element, but the XSLT 2.0 WD also notes "It is possible for two conforming implementations not to convert a number to exactly the same string."
A weak recommendation to read a Unicode TR does not do much for a tool developer. However, mandating a large number of languages whose sort orders and number-to-string conversion conventions must be supported by all XSLT processors simply invites non-conformance by tool developers not implementing onerous requirements, and also doesn't do anything for languages not on the official list.
The W3C I18N WG could, however, produce guidelines about collation and number-to-string conversion that provides either the necessary information or pointers to where to find the necessary information so that tool developers face fewer hurdles when they want to support collation and numbering for languages other than English.
The W3C I18N WG should engage in more coordination/joint activities with I18N/Localisation groups in other organizations, such as the newly formed XLIFF TC in OASIS.
For example, "Unicode in XML and other Markup Languages," which is both a W3C Note and Unicode's UTR #20, shows effective coordination between the W3C and another organization, since the Note/UTR provides a single source of information for both those primarily interested in XML and those primarily interested in extending their existing Unicode support to also cover XML.
Joint activities and joint publications such as the Note/UTR provide a unified message, make it easier for developers to find the right information quickly, and show the work of the W3C I18N WG to developers who might not have otherwise found it.