# The following was supposedly scribed by # Bruno Postle # on Wednesday 25 August 2004 06:46 am: >> I'm afraid rhizopod isn't extensible. > >How are we supposed to play with this stuff? =A0Specifications have to >evolve under real usage conditions - They are never right first time. One way is to "subclass the spec" as I said earlier. I don't think I'll ge= t=20 rhizopod perfect on the first go, but things like polymorphic inheritance a= nd=20 surprise changes to entity versions don't fit the rhizopod model of=20 simplicity. >> But 'grep' wasn't compiled with my library. > >All files will be opened with the YAML library, if the 'version' >attribute is always there (along with 'type'), there is no need to >use grep. But I *want* to be able to use grep without having to teach it about versio= ns. =20 To put it another way: I need to be _able_ to use grep. =2D-Eric =2D-=20 "...our schools have been scientifically designed to=20 prevent overeducation from happening." --William Troy Harris