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