gitpan / string-downgrade Goto Github PK

I don't yet have a good name for this package.  String::Amharic::Simplify
might be more immediately understandable.  However, "simplification" is
generally considered a good thing, something desirable, but this would
give quite the wrong impression.  String::Amharic::Pencanonicalize was a
working name for a while, now its String::Amharic::Downgrade which seems
to have the best tone.

Under the "three levels of Amharic spelling" theory, this package will
take a canonical word (level 1) and generate level two words (the level
of popular use).  Hence "pencanonical", though think spherically and the
"pen" part of it spans the breadth of level two.  Its a little easier to
think of a simple "downgrade" vs debating the width of "pen".

The package is useful for some problems, it will produce orthographically
"legal" simplification and avoids improbable naive simplifications.
Text::Metaphone::Amharic of course over simplifies as it addresses a
different problem.  So while not to promote level 2 orthographies, for some
instances it is useful to generate level 2 renderings given a canonical
form.  The package is intended to be a labor saving device and is by no
means perfect.

You *must* start with the canonical spelling of a word as only downgrades
can occur.  Starting with a near canonical form and downgrading will generate
a shorter word list than you would have starting from the top.