reduce-algebra/contrib/LuFy/README.lufy

              License-related comments about the files here.

The file "lufy.red" is made by applying a series of mechanical
transformations starting from tex.web, which is the Original Knuth
literate form of the TeX program source. The exact details of these
transformations are documented precisely by the code here, but clearly
many people will consider lightly commented code not to be as clear
an explanation as would be desirable even if it is (by definition)
pretty complete.

p2l.l is the flex input that defines a Pascal-compatible lexical analyser.
p2l.ypp is a bison input file describing a Pascal parser. I added semantic
    actions to an existing grammar so that some form of parse tree gets
    created. Each individual semantic action should be simple enough to
    understand.
p2l.cpp contains two segments of code. One lot is a set of constructors
    for parse tree nodes, which are built in a really naive manner. The
    second is code that traverses a parse tree printing it in a fully
    parenthesised notation that will be acceptible to a Lisp system.

The code in all the above is supposed to preserve all information present
in Pascal source (which is itself obtained by using "tangle" in tex.web).

p2l.red is Reduce code that reads in the parenthesised parse tree and
   rearranges it so that it ends up as something rather like executable
   Lisp code. In doing this it discards Pascal type information, but in
   general it is mostly just performing fairly direct re-writes of the
   structure of the tree. For instance in the bison parser most lists are
   accumulated in a left-associated form, as in (((a b) c) d), and p2l.red
   will typically change that to be (a b c d). It also normalises the names
   at the head of segments of the parse tree to match function names in
   the Standards Lisp that Reduce uses. The output from p2l.red is lufy.red.

A number of Pascal features do not translate into Lisp very easily, and so
the code in lufy.red is not directly usable. To make it possible to run it
a prolog file is provided that defines some of the functions present in
in in a way that is compatible with native Lisp, and then flags the names
of those functions as 'LOSE. That is a Reduce feature which permits the
definition from the prolog to take the place on the one in lufy.red.

As regards the rights associated with the TeX source, lufy.red is clearly
derived from tex.web, but its name is not anything that can be readily
confused with TeX and that should satisfy the requirements of the Knuth
License. The detailed documentation of how it is derived available by
inspecting the code used to make it, and the fact that a prolog file of
overrides is used rather than hand editing of the transformed file are
schemes intended to fit in with the spirit as well as the letter of TeXs
requirements.