1 _____________________________________________________________________________
4 - If a top-level rule has labels but no head-tag, like this
6 then infer the name of the rule it belongs to
8 create( $c:{...}, class ) =
9 return create($c:{...})
11 create( h:{...}, class ) =
15 create( _:{...}, String) = treat as char[]
16 create( _:{...}, c[] ) = { create(.,c), create(.,c), ... }
21 - better ambiguity debugging tools / visualization
23 - ParseFailed, GSS, Walk, Parser, Sequence, Forest
25 - Fix the metagrammar (really?)
26 - evil problems with (x y? z /ws)
31 ______________________________________________________________________________
34 - finalize metagrammar and rdp-op's
39 - RFC2822 (email message/headers)
40 - clean up the whole Walk situation (?)
42 - what if Tree<> could unwrap itself?
45 ______________________________________________________________________________
48 - serialization of parse tables
50 - "ambiguity modulo dropped fragments"?
51 - can this be checked statically?
52 - eliminated statically?
54 - substring parsing for better error messages
56 - right now I can only lift the last child in a forest... begs
57 the question of what the right representation for Forests is
58 if we need to be able to do lift operations on it.
61 - "Regular Right Part" grammars (NP Chapman, etc)
62 - Attribute unification
64 - inference of rejections for literals
65 - "prefer whitespace higher up" (?)
67 - Labeled edges on trees (associate a label with each slot in the
68 child array in Forest.Body? might make equality tough) --
69 equivalent to Feature Structures. Colon-labeling.
71 ______________________________________________________________________________
74 - Partly-Linear-PATR? (O(n^6) unification grammar)
76 - Implement a k-token peek buffer (for each state, see if it "dead
77 ends" during the next k Phases based solely on state -- ignoring
80 - Arrange for the SPPF corresponding to dropped subtrees to never be
81 generated (or merged, etc)
83 - Is there any way we can avoid creating a GSS.Node instance for
84 nodes which are transient in the sense that they have only one
87 - Re-read Rekers, particularly the stuff on optimal sharing
89 - Isolate the Element objects from Parse.Table/GSS so we can move
92 - consider allowing a Forest.Body to represent some other Tree whose
93 Body's should be [recursively] considered part of this Forest.
95 - perhaps not: right now we have a nice situation where
96 Forest.Ref instances become immutable once iterator()ed. This
97 also gives us a strong place to to culling with the certainty
98 that we won't throw out a Body which would later be salvaged
99 by some yet-to-be-added dependency.
101 - Figure out if there is a way to:
103 - allow unwrapping of children other than the very last one.
105 - fold repetitions into an array form in Forest, before
106 conversion to Tree. The major problem here is that multiple
107 tree-arrays are possible, all of different lengths. Worse,
108 even if they're all the same length, not all elements belong
109 in the same "possibility vector" as all others. You
110 essentially need a GSS to represent the array, which perhaps
111 is what the unfolded form was in the first place.
113 - Wikipedia grammar (needs to be both lexerless and boolean)
116 => Ordered Choice (";" operator)
118 - bring back in parse-table phase resolution of precedence (just
119 like associativity). This can be inferred from the use of ">"
120 when the rules are in one of these special forms:
131 where "_" is anything and "E" is the defining nonterminal.
132 Essentially what we're looking for is the situation where the
133 leftmost portion of one rule produces another rule, and the
134 rightmost portion of the latter produces the former.
136 I'm not 100% certain that this is as "strong" as the prefer/avoid
137 form (try to prove this, you probably can), but it's "what people
138 intend" most of the time.
140 - implement Johnstone's algorithm for "reduced, resolved LR
141 tables" to eliminate superfluous reductions on
144 ______________________________________________________________________________
147 - Rekers & Koorn note that GLR Substring Parsing can be used to do
148 really elegant and generalized "autocompletion".
151 ______________________________________________________________________________
154 - Incremental parse table construction
155 - "lazy GLR" and "lazy trees" -> language with first-class CF matching
156 - perhaps linear boolean grammars instead? (linear time, quad space)
157 - Forest parsing => chained parsers
158 - unification parsing, attributes, etc
160 - Take another stab at maximal-match? Nonterminal not-followed-by is
162 - Error recovery based on substring parsing