Avik Sengupta wrote: >>Go for it. >> >Sure. > >>I'd prefer if you wait until after we get the branch though so that I >>can commit. >> > >yeah, thats what i was planning. Given that i am situated between glen and you >timewise, it should be done by the time you wake up tomorrow :) > You may want to wait until after I commit what I've done so far before you make this correction.
> >>There will also likely be a "FunctionPtg" abstract class. >> > >I guess. Some sort of factory as well, probably? > Yeah, but I don't want to think that far ahead. The first thing we need to do is figure out what we need to do to tokenize, RPN and the reverse. Once we figure that out, we'll create a good object model to do it in. I think we need prototypical code and or "tracer" code before we start guessing at the full object model. Thats just my opinion. Once we figure out our approach, I expect that a lot of refactoring will take place. -Andy > >Quoting "Andrew C. Oliver" <[EMAIL PROTECTED]>: > >>Go for it. In general, I don't want to focus on such detail oriented >>things just yet. Figure we'll change this around a couple times. >> >>Locally, my OperationPtg is an Abstract class inheriting from the PTG >>class. There will also likely be a "FunctionPtg" abstract class. >> >>I'd prefer if you wait until after we get the branch though so that I >>can commit. (Should happen in about 12 hours or so when Glen is awake.) >> >>Avik Sengupta wrote: >> >>>Can i suggest that we make Ptg an interface, and have an AbstractPtg class. >>> >>>Seems rather incongrous to me that OperationPtg, being a subset of ptg, >>> >>should >> >>>be an interface, while Ptg is a class. >>> >>>If there are no objections, i will make a patch. >>> >>>Regards >>>- >>>Avik >>> >> >> >> > > >
