Re: Reversible sound change applier

Hi!

Jamie Norrish writes:
> Henrik Theiling writes:
>
> [IPA Zounds]
>  > But it seems a bit tedious to write in phonetics instead of being
>  > able to define phonemes.  But maybe I missed something.
>
> You can in fact create a script file which maps between IPA
> characters/clusters and whatever character/cluster you want.
>
>  > Does it support user-defined syllable structure and/or syllable
>  > constraints?
>
> In a sense - you can specify syllable boundaries in rules (using
> suprasegmental characters which carry that feature, such as stress
> marks, or by the feature itself), and in the reverse applier the
> possible antecedent words are limited by (optional) syllable structure
> definitions.
Ah, ok, then that's both possible.  Nice! :-)

>...
>  > Global and local boolean flags?
>
> Nope - what is your use case?
In the way you can mark for dialects, my script allows marking rules
for global flags which you can set to true or false before running a
conversion, switching on and off thusly constrained rules.  I call the
global flags 'options'.  Declaration syntax is as follows:

option Option1 = 0

Usage:

rule "Some rule"
   option Option1
   ...

The local flags are similar in usage, but setting them is by a
matching rule.  This way, you can enable/disable rules by matching of
other rules.  I implemented the syncope in Old Norse this way, which
happens either in the last syllable (by phonological constraints), or
if not there, then in the penultimate syllable.  Definition syntax
is as follows:

rule "blah 1"
    syllable last
    V  > 0 / somecontext    => syncope_happened


rule "blah 2"
    syllable penultimate
    unless syncope_happened
    V  > 0 / someothercontext

**Henrik