1 _____________________________________________________________________________
4 - Topology crap is kinda messed up
5 - Atom should be a topology, shouldn't it?
7 - needs/hates/follow API ugliness
9 - do Forest/Tree still need a Region?
10 - reconsider the degree of genericization
11 - GraphViz stuff pollutes the API...
12 - Forest needs a "manual access" API
13 - the unwrap bit in Forest makes it really hard to expose an API for forests
15 - evil problems with (x y? z /ws)
16 - ParseFailed, GSS, Walk, Parser, Sequence, Forest
20 - grammar highlighting?
21 - comment indentation vs block indentation?
23 - recursive { { foo } }
25 ______________________________________________________________________________
28 - finalize metagrammar and rdp-op's
33 - RFC2822 (email message/headers)
34 - clean up the whole Walk situation (?)
37 ______________________________________________________________________________
40 - serialization of parse tables
42 - "ambiguity modulo dropped fragments"?
43 - can this be checked statically?
44 - eliminated statically?
46 - substring parsing for better error messages
49 - "Regular Right Part" grammars (NP Chapman, etc)
50 - Attribute unification
52 - inference of rejections for literals
53 - "prefer whitespace higher up" (?)
55 - Labeled edges on trees (associate a label with each slot in the
56 child array in Forest.Body? might make equality tough) --
57 equivalent to Feature Structures. Colon-labeling.
59 ______________________________________________________________________________
62 - Partly-Linear-PATR? (O(n^6) unification grammar)
64 - Implement a k-token peek buffer (for each state, see if it "dead
65 ends" during the next k Phases based solely on state -- ignoring
68 - Arrange for the SPPF corresponding to dropped subtrees to never be
69 generated (or merged, etc)
71 - Is there any way we can avoid creating a GSS.Node instance for
72 nodes which are transient in the sense that they have only one
75 - Re-read Rekers, particularly the stuff on optimal sharing
77 - Isolate the Element objects from Parse.Table/GSS so we can move
80 - consider allowing a Forest.Body to represent some other Tree whose
81 Body's should be [recursively] considered part of this Forest.
83 - perhaps not: right now we have a nice situation where
84 Forest.Ref instances become immutable once iterator()ed. This
85 also gives us a strong place to to culling with the certainty
86 that we won't throw out a Body which would later be salvaged
87 by some yet-to-be-added dependency.
89 - Figure out if there is a way to:
91 - allow unwrapping of children other than the very last one.
93 - fold repetitions into an array form in Forest, before
94 conversion to Tree. The major problem here is that multiple
95 tree-arrays are possible, all of different lengths. Worse,
96 even if they're all the same length, not all elements belong
97 in the same "possibility vector" as all others. You
98 essentially need a GSS to represent the array, which perhaps
99 is what the unfolded form was in the first place.
101 - Wikipedia grammar (needs to be both lexerless and boolean)
104 => Ordered Choice (";" operator)
106 - bring back in parse-table phase resolution of precedence (just
107 like associativity). This can be inferred from the use of ">"
108 when the rules are in one of these special forms:
119 where "_" is anything and "E" is the defining nonterminal.
120 Essentially what we're looking for is the situation where the
121 leftmost portion of one rule produces another rule, and the
122 rightmost portion of the latter produces the former.
124 I'm not 100% certain that this is as "strong" as the prefer/avoid
125 form (try to prove this, you probably can), but it's "what people
126 intend" most of the time.
128 - implement Johnstone's algorithm for "reduced, resolved LR
129 tables" to eliminate superfluous reductions on
132 ______________________________________________________________________________
135 - Rekers & Koorn note that GLR Substring Parsing can be used to do
136 really elegant and generalized "autocompletion".
139 ______________________________________________________________________________
142 - Incremental parse table construction
143 - "lazy GLR" and "lazy trees" -> language with first-class CF matching
144 - perhaps linear boolean grammars instead? (linear time, quad space)
145 - Forest parsing => chained parsers
146 - unification parsing, attributes, etc
148 - Take another stab at maximal-match? Nonterminal not-followed-by is
150 - Error recovery based on substring parsing