1 _____________________________________________________________________________
4 - I still don't like Atom.Infer and Atom.Invert...
6 - Fix the metagrammar (really?)
8 - decent/better error messages
9 - fix the location stuff, it's broken
14 ______________________________________________________________________________
17 - finalize metagrammar and rdp-op's
22 - RFC2822 (email message/headers)
23 - clean up the whole Walk situation (?)
26 ______________________________________________________________________________
29 - serialization of parse tables
31 - "ambiguity modulo dropped fragments"?
32 - can this be checked statically?
33 - eliminated statically?
35 - substring parsing for better error messages
37 - right now I can only lift the last child in a forest... begs
38 the question of what the right representation for Forests is
39 if we need to be able to do lift operations on it.
42 - "Regular Right Part" grammars (NP Chapman, etc)
43 - Attribute unification
45 - inference of rejections for literals
46 - "prefer whitespace higher up" (?)
48 - Labeled edges on trees (associate a label with each slot in the
49 child array in Forest.Body? might make equality tough) --
50 equivalent to Feature Structures. Colon-labeling.
52 ______________________________________________________________________________
55 - Partly-Linear-PATR? (O(n^6) unification grammar)
57 - Implement a k-token peek buffer (for each state, see if it "dead
58 ends" during the next k Phases based solely on state -- ignoring
61 - Arrange for the SPPF corresponding to dropped subtrees to never be
62 generated (or merged, etc)
64 - Is there any way we can avoid creating a GSS.Node instance for
65 nodes which are transient in the sense that they have only one
68 - Re-read Rekers, particularly the stuff on optimal sharing
70 - Isolate the Element objects from Parse.Table/GSS so we can move
73 - consider allowing a Forest.Body to represent some other Tree whose
74 Body's should be [recursively] considered part of this Forest.
76 - perhaps not: right now we have a nice situation where
77 Forest.Ref instances become immutable once iterator()ed. This
78 also gives us a strong place to to culling with the certainty
79 that we won't throw out a Body which would later be salvaged
80 by some yet-to-be-added dependency.
82 - Figure out if there is a way to:
84 - allow unwrapping of children other than the very last one.
86 - fold repetitions into an array form in Forest, before
87 conversion to Tree. The major problem here is that multiple
88 tree-arrays are possible, all of different lengths. Worse,
89 even if they're all the same length, not all elements belong
90 in the same "possibility vector" as all others. You
91 essentially need a GSS to represent the array, which perhaps
92 is what the unfolded form was in the first place.
94 - Wikipedia grammar (needs to be both lexerless and boolean)
97 => Ordered Choice (";" operator)
99 - bring back in parse-table phase resolution of precedence (just
100 like associativity). This can be inferred from the use of ">"
101 when the rules are in one of these special forms:
112 where "_" is anything and "E" is the defining nonterminal.
113 Essentially what we're looking for is the situation where the
114 leftmost portion of one rule produces another rule, and the
115 rightmost portion of the latter produces the former.
117 I'm not 100% certain that this is as "strong" as the prefer/avoid
118 form (try to prove this, you probably can), but it's "what people
119 intend" most of the time.
121 - implement Johnstone's algorithm for "reduced, resolved LR
122 tables" to eliminate superfluous reductions on
125 ______________________________________________________________________________
128 - Rekers & Koorn note that GLR Substring Parsing can be used to do
129 really elegant and generalized "autocompletion".
132 ______________________________________________________________________________
135 - Incremental parse table construction
136 - "lazy GLR" and "lazy trees" -> language with first-class CF matching
137 - perhaps linear boolean grammars instead? (linear time, quad space)
138 - Forest parsing => chained parsers
139 - unification parsing, attributes, etc
141 - Take another stab at maximal-match? Nonterminal not-followed-by is
143 - Error recovery based on substring parsing