About package naming and the context tokenizer, I was quite puzzled as to why no one had so far complained about the compilation issues in the Git Archive which I noticed.
The issue is that a bunch of the ctakes files refer to a package *org.apache.ctakes.* *contexttokenizer/...* when its contents actually live in the folder *org/apache/ctakes/context/**tokenizer/....* I did some research and discovered something that I hadn't known. Apparently the Java spec suggests but doesn't enforce that package names and folder structure should mirror each other. While Eclipse enforces it, some other build environments may not. This was reported to the Eclipse team years ago and was assigned "wont-fix" status. I think I agree with that decision. Since Java's consistency is one of its great virtues, with class names required to mirror file names, why allow fuzzy folder placement of sources? In the case of the Git archive for ctakes, the folders are already logical and "correct", but in some files the package names and imports for the *context.tokenizer* are mismatching. Since I do use Eclipse, I know that the context.tokenizer is the only instance of this issue. Would anyone mind if I corrected the package names and references to match the folders? Peter
