Open oerich opened 12 years ago
Possibly you should also take a look at existing implementations in Apache lucence or some other nlp processing framework. However I think the goal of the project has to be defined first to know what should be reached, before investigating any time in refactoring. a. Should this be a meta framework, like for example the springframework, to provide an API to be able to work with different nlp frameworks exchangeable, or b. should it provide its own nlp implementations, or c. should it be a mix of a. and b?
After the goal had been cleared some quality goals/requirements should be defined, for example in which context the library should be used "sandbox" projects only or large nlp processing programs.
I think to clearify the goals/requirements is very important to have a clear direction for refactoring.
The factory is mixed up with the implementation. This has to be cleaned up. Also, there are too many alternatives flying around. We only need one clean mechanism to filter stopwords. If we need alternatives, these have to be clearly named and consistent in how they should be initiated and used.