- From: Tab Atkins Jr. <jackalmage@gmail.com>
- Date: Sun, 22 Jan 2012 15:16:18 -0800
- To: Brian Manthos <brianman@microsoft.com>
- Cc: www-style list <www-style@w3.org>
On Sun, Jan 22, 2012 at 2:53 PM, Brian Manthos <brianman@microsoft.com> wrote: > TR > [ > [ top | bottom ] > | > [ <percentage> | <length> | left | center | right ] > [ <percentage> | <length> | top | center | bottom ]? > | > [ center | [ left | right ] [ <percentage> | <length> ]? ] && > [ center | [ top | bottom ] [ <percentage> | <length> ]? ] > ] > > > ED (modified as per Tab below) > [ > [ left | center | right | top | bottom | <percentage> | <length> ] > | > [ left | center | right | <percentage> | <length> ] > [ top | center | bottom | <percentage> | <length> ] > | > [ center | [ left | right ] [ <percentage> | <length> ]? ] && > [ center | [ top | bottom ] [ <percentage> | <length> ]? ] > ] > > > Marking the sub-clauses... > > TR > (A) [ top | bottom ] > (B) [ <percentage> | <length> | left | center | right ] > [ <percentage> | <length> | top | center | bottom ]? > (C) [ center | [ left | right ] [ <percentage> | <length> ]? ] && > [ center | [ top | bottom ] [ <percentage> | <length> ]? ] > > ED(m) > (A) [ left | center | right | top | bottom | <percentage> | <length> ] > (B) [ left | center | right | <percentage> | <length> ] > [ top | center | bottom | <percentage> | <length> ] > (C) [ center | [ left | right ] [ <percentage> | <length> ]? ] && > [ center | [ top | bottom ] [ <percentage> | <length> ]? ] > > > TRC and EDmC match, so let's ignore those for the moment. > > > Unless I'm misreading it, (A) has become more complicated just so 1 character can be removed from (B). > > > Tab, why is the ED version of (A)+(B) considered an improvement over the TR version of (A)+(B)? Because it more closely matches the actual productions. In the ED version, the (A) clause covers all the 1-token productions, the (B) clause covers most of the 2-token productions (with the exception being [top|bottom] [left|right]), and the (C) clause covers everything else. On the other hand, in the TR version, the (A) clause covers two of the 1-token productions, while (B) covers the rest of the 1-token productions and most of the 2-token productions, and then (C) covers everything else. > Shouldn't it be demonstrably better to be worth changing? The productions are exactly equivalent, so this amounts to an editorial change. That said, I think the suggested version is easier to read and understand. ~TJ
Received on Sunday, 22 January 2012 23:17:09 UTC