This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
Section 3.2.6 (FTLanguageOption) Appendix C (Static Context) Appendix I (Checklist of Implementation-Defined) language="none" is not a sensible choice. Tokenization is intrinsically language-dependent and full-text search is intrinsically dependent on tokenization. I believe we should drop the language=none and require the static context to define a specific (implementation-defined) default language.
This overstates the rationale somewhat: there are many languages for which a simple whitespace/punctuation-delimited tokenization does the job. However, there are also languages where tokenization is entirely language-dependent. Further, stemming is certainly entirely language dependent, so at the very least the spec needs to say what a stemmed query with language=none means. In general, we still prefer that language=none be dropped, so that those kinds of messy questions can be avoided entirely.
The WG decided to change the specification: The default language is specified in the static context. Remove None. Add "in an implementation-defined way" at the end of the sentence: The "language" option influences tokenization, stemming, and stop words. 3rd para DELETE 1) either 2), or be the value "none". DELETE the 4th para: If the language "none" option is specified, no language selected. 6th para change to: The default language is specified in the static context. Annex C: 1) correct FTLanguageOption row: Instead of "no language is selected" use" implementation-defined" 2) correct last column eliminate "or none".