2 ______________________________________________________________________________
5 - switch maximal to not-followed-by (~/~)
7 - should Union.add() be there?
8 - should Atom.top() be there?
10 - fix the location stuff, it's broken
11 - decent/better error messages
17 - RFC2822 (email message/headers)
20 ______________________________________________________________________________
23 - clean up the whole Walk situation
25 - cleaner solution to "maximal"?
28 - right now I can only lift the last child in a forest... begs
29 the question of what the right representation for Forests is
30 if we need to be able to do lift operations on it.
33 - "Regular Right Part" grammars (NP Chapman, etc)
34 - Attribute unification
36 ______________________________________________________________________________
39 - Implement a k-token peek buffer (for each state, see if it "dead
40 ends" during the next k Phases based solely on state -- ignoring
43 - Arrange for the SPPF corresponding to dropped subtrees to never be
44 generated (or merged, etc)
46 - Is there any way we can avoid creating a GSS.Node instance for
47 nodes which are transient in the sense that they have only one
50 - Implement "GLR syntactic predicates" -- the ability to do
51 arbitrary lookahead (ie "followed-by" and "not-followed-by" for
52 arbitrary patterns). This enables generalized longest-match and
53 lets us drop the Maximal hack.
55 - Re-read Rekers, particularly the stuff on optimal sharing
57 - Isolate the Element objects from Parse.Table/GSS so we can move
60 - consider allowing a Forest.Body to represent some other Tree whose
61 Body's should be [recursively] considered part of this Forest.
63 - perhaps not: right now we have a nice situation where
64 Forest.Ref instances become immutable once iterator()ed. This
65 also gives us a strong place to to culling with the certainty
66 that we won't throw out a Body which would later be salvaged
67 by some yet-to-be-added dependency.
69 - Figure out if there is a way to:
71 - allow unwrapping of children other than the very last one.
73 - fold repetitions into an array form in Forest, before
74 conversion to Tree. The major problem here is that multiple
75 tree-arrays are possible, all of different lengths. Worse,
76 even if they're all the same length, not all elements belong
77 in the same "possibility vector" as all others. You
78 essentially need a GSS to represent the array, which perhaps
79 is what the unfolded form was in the first place.
81 - Wikipedia grammar (needs to be both lexerless and boolean)
84 => Ordered Choice (";" operator)
86 - bring back in parse-table phase resolution of precedence (just
87 like associativity). This can be inferred from the use of ">"
88 when the rules are in one of these special forms:
99 where "_" is anything and "E" is the defining nonterminal.
100 Essentially what we're looking for is the situation where the
101 leftmost portion of one rule produces another rule, and the
102 rightmost portion of the latter produces the former.
104 I'm not 100% certain that this is as "strong" as the prefer/avoid
105 form (try to prove this, you probably can), but it's "what people
106 intend" most of the time.
108 - implement Johnstone's algorithm for "reduced, resolved LR
109 tables" to eliminate superfluous reductions on