Revision history for Lingua-EN-Titlecase
0.04 Mon Aug 20 22:52:00 PDT 2007
- Totally changed the guts. Now a lexer queue drives it instead of
a single pass set of regexes. Looks a little spaghetti-ish but I
think it'll clean up nice and allow for a lot of easy add-ons
like dictionary hooks.
- Added List::Util to get "first" token in queue which is wordish
without a placeholding switch.
- Cleaned up wordish/wc regex.
- Mixed case terms with leading lowercase are now correctly
preserved if they are below the built-in or user-set threshold.
- Raised uppercase threshold from 0.9 to 0.95.
- Lowered mixedcase threshold from 0.3 to 0.25 based on some test
strings which show a pretty good breakpoint around 0.15.
- Removed unused accessors.
- A few small logic improvements in places like counting the real
length of the string for computing mixedcase and uppercase
ratios.
0.03 Mon Aug 20 14:46:17 PDT 2007
- Changed the "apostrophe" to a single [[:punct:]]. This alters
behavior quite a bit. Moo-ish and search.cpan.org are now one
word. They were two and three before. Mixed case matches
similarly differently too.
- I think to deal with stuff like "iPods on eBay!" the titlecase
method, and general handling, will have to be token-based. It
gets too messy, too quickly otherwise.
0.02 Mon Aug 20 01:12:17 PDT 2007
- Finding fault so quick: POD example updated with correct output.
0.01 Sun Aug 19 23:36:33 PDT 2007
- Initial release. Champagne for everyone but the dogs! Sparkling
wine for the dogs!