This looks good. It is low-risk, since it changes no existing classes, adding new classes in a new package. It looks as though you've implemented the right things (a stemming filter, and an analyzer that plugs this together with a stop list and tokenizer). I don't speak Dutch, so I cannot assess the quality of these.
My concerns before this is checked in are that: 1. It compiles. I have not tested this yet, and we don't want to break the build. 2. More Javadoc is added. My rule of thumb is that there should be no empty boxes in the generated javadoc, i.e., every publicly visible class, method and field must be documented. Also, every package should have a package.html with at least a one-line description of what is in the package. 3. Fewer classes and methods are public. Do the Rule, RuleVector and DutchStemmer classes need to be public? Probably little more than the core classes (DutchStemFilter DutchAnalyzer) need to be public. Doug > -----Original Message----- > From: Maurits van Wijland [mailto:[EMAIL PROTECTED]] > Sent: Wednesday, October 10, 2001 12:48 AM > To: [EMAIL PROTECTED] > Cc: [EMAIL PROTECTED]; Doug Cutting > Subject: Dutch Stemmer > > > Doug, > > Here is my first version of the dutch stemmer. > Doug could you have a look at this? > > kind regards, > > Maurits > > >
