On 7/7/07, John Randall <[EMAIL PROTECTED]> wrote:
J has a published standard (the Dictionary) with a few hints as to what should be implemented efficiently (the targets for "special code" given in the release notes). However, there is only one closed-source, but noncommercial implementation.
There's also source code for an earlier version of J, visible from http://www.math.uwaterloo.ca/apl_archives/j/Welcome.html (near the bottom of the page). The problem, I think, is an issue of "derivative works". Under copyright law, the copyright holder "owns" works based on the ideas of the original publication. In practical terms, this protection lasts indefinitely. While, in principle, mathematical concepts are not supposed to be copyrightable, most people are uncertain what that means, legally, in the context of computers. Thus, even though Roger has stated that it's OK for people to write implementations of J based on the dictionary, since he is not the copyright holder (ISI is), it's still not very clear what is and is not legal. For example, if I were writing an implementation of J, I would prefer to incorporate the relevant specifications as embedded comments, if not user accessible documentation. But I doubt either would be legal. And, while I personally am comfortable with the idea that if I were to write and release a J implementation (that did not incorporate the actual text of the specification in the body of code) that ISI would not take legal action against me, I'm not sure that other people have the same level of comfort. -- Raul ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
