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