On 9 December 2012 17:25, Don Guinn <[email protected]> wrote: > The point is that the context must be determined before an expression is > evaluated. The rules for evaluation of an expression should be context > independent.
There is always a context. One cannot escape from that, except in very simple domains. School algebra seems to be such a domain, or at least possible context dependencies are minimal there. In J, on the contrary, even parsing is context-dependent. In order to parse an expression, one has to know which of the names in it designate nouns, which are verbs, etc. Without it, one does not know how to form sub-expressions. This is context dependence taken to the extreme. Earlier I gave an example of how a sequence of two sub- expressions in J can have ten different interpretations, depending on the types of the sub-expressions. ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
