laurentg / motscroises Goto Github PK

Everything is said in the comment... favor words according to some weight. The weight could come from different sources:

• A dictionnary defining a weight for each word
• Multiple dictionnaries with a weight for each, applying to every word
• ...

Currently the algorithm does not take into account the removed number of black cells to determine which word to remove first. This has some impact so this should be taken into account.

Score some words from sub-patterns even when the maximum word count to score is reached using only the full pattern.
Currently when the max word count to analyze is reached during full pattern analysis, we break the loop, without looking at some sub-patterns. We should always look down at some sub-patterns, as some may lead to better results than full pattern. This behavior also strongly favor very long words, which may be sub-optimal, especially if we use word weighting (see issue #1).