|
Specifically...
* Bogowin was only being used for English-like dictionaries.
I suppose the bogowin numbers might change a bit from
dictionary to dictionary, but I think it's quite plausible that
the numbers are similar for all dictionaries.
* This fixes bogowin always returning 100 or 0.
* Make the strategy code more fine-grained, so that if it
has some strategy files and not others, it will use what it
has. JKB recommended a long time ago that I add a generic
worths file to give the blank a high worth, but it wasn't being
used because most languages didn't also have a syn2 or
a superleaves. Now it will. Also, the vowel-consonant
balance computations don't use any strategy files at all,
but they were also being skipped if you didn't have every
single strategy file. Oops.
* The strategy computations would sometimes do
unexpected things if the leaves weren't alphabetized.
They now look at the alphabetized version of the leave.
* A couple more conversions to ranged-for loops.
|
|
So, basically, the dictionary is queried to see if it's
"English-like". Which is to say that it uses the same
alphabet. If so, then it will fall back to use twl06
strategies, minus the superleaves (the so-called
"default_english"). Otherwise, it uses "default" strategies,
which right now is simply a worths file that sets the
blank to be 30.
It's not great. But better than before.
|
