James Edward Gray II wrote: > On Nov 28, 2005, at 2:10 PM, Leslie Viljoen wrote: > >> Comments very welcome! > > > I haven't had a chance to dig through your code yet, but I just > wanted to stress that you look through all the solutions to that quiz > if you haven't already. A lot of good ideas in there. Jim Menard > also built a RADS library, for example... I saw that, and I looked at many of the solutions. AFAIR Jim's parser allowed running arbitrary methods in Object - so I wanted to take that out, and also bend a solution to work a bit more like TADS. In my program the "kick" verb consults the object (bucket) to see if it's "kickable". This is very similar to how TADS works with objects. The real challenge will be the parser - in fact, that might be a good quiz too. TADS has thousands of lines dedicated to tokenising and parsing english - breaking sentences into phrases and phrases into adverb + (noun phrase) where (noun phrase is adverb + noun-phrase). Almost a regex pattern matcher for word types. It has a ranking system where it tries to figure out the most likely thing you are trying to say instead of just taking the first possible meaning. I'm still going through the TADS3 parsing code trying to understand it. Les