This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
The weight option is defined by the following production: [144] FTSelection ::= FTOr FTPosFilter* ("weight" RangeExpr)? This means that in this expression the weight applies to the FtAnd expression, not the term "bar": "foo" ftand "bar" weight 2 This is equivalent to: ("foo" ftand "bar") weight 2 No to: "foo" ftand ("bar" weight 2) Now, my feeling is that when writing the 1st expression, a programmer would probably mean the second form: "foo" ftand ("bar" weight 2) In other terms, this syntax seems misleading to me. A related problem is that it is possible to specify a weight twice on the same expression: ("foo" ftand "bar" weight 2) weight 1 Here both weight options apply to the FtAnd. Suggestion: move the "weight" option to the level of FTPrimaryWithOptions: [149] FTPrimaryWithOptions ::= FTPrimary FTMatchOptions? ("weight" RangeExpr)? [144] FTSelection ::= FTOr FTPosFilter*
Xavier, The Full Text Task Force at its 10/30/2008 face-to-face meeting agreed to change the grammar for weights as you proposed to: [149] FTPrimaryWithOptions ::= FTPrimary FTMatchOptions? ("weight" RangeExpr)? [144] FTSelection ::= FTOr FTPosFilter* Thank you for making this suggestion. The change will appear in the next published version of specification. We also discussed the related problem that you raised -- that it is possible to specify a weight twice on the same expression. Because the effect of weights on the resulting score is implementation-dependent, we decided to leave this detail implement-dependent as well. Pat Case, Member, Full Text Task Force
Xavier, If you approve of the response to the bug noted above, please close the bug. Pat Case, Member, Full Text Task Force