Loading source files for package notio...
Loading source files for package notio.translators...
Loading source files for package notio.test...
Constructing Javadoc information...
/home/finnegan/code/src/notio/translators/LFParser.java:6: Class com.metamata.parse.ParseException not found in import.
import com.metamata.parse.ParseException;
^
/home/finnegan/code/src/notio/translators/CGIFParserImpl.java:7: Package com.metamata.parse not found in import.
import com.metamata.parse.*;
^
/home/finnegan/code/src/notio/translators/CGIFParserImpl.java:241: Superclass com.metamata.parse.Scanner of inner class notio.translators.CGIFParserImpl. __jjScanner not found.
public class __jjScanner extends com.metamata.parse.Scanner {
^
/home/finnegan/code/src/notio/translators/CGIFParserImpl.java:311: Superclass com.metamata.parse.ScanError of inner class notio.translators.CGIFParserImpl. __jjScanner.TokenMgrError not found.
class TokenMgrError extends com.metamata.parse.ScanError
^
/home/finnegan/code/src/notio/translators/CGIFParser.java:6: Class com.metamata.parse.ParseException not found in import.
import com.metamata.parse.ParseException;
^
List of @idea Tags
List of @idea Tags
Package Index
Class Comments
- If we switch to JDK 1.2, consider allowing weak references to individuals.
Class Comments
- Add relatedBy() method.
Class Comments
- Add OutermostContext() or TopLevelGraph()? This would allow for automatic
clearing of def. label tables, etc.
Class Comments
- Add isolateGraph() method to remove coref links to other graphs but
not internal links?
- How about maintaining a 'connected subgraph' ID for each node in the graph,
which is set when it is added. For concepts, check relators for IDs. For relations,
check arguments for IDs. When two IDs are found the same, somehow identify them
with each other. Perhaps bitfields: bit for each new connected subgraph found.
Also use a set of masks to represent the real groups.
If two are later found to be connected, the masks are combined but the bitfield on
each node can remain the same. Drag if we have to recompute the whole thing after
disconnecting something. Better to change everyone you are connected to. Or, could
create an int and point to it whenever a new subset is found. If two different ints
are found to be connected, one of them simply changes value to that of the other.
- What about "graph references". Could implement the same interface as Graph, but
actual be a proxy class for most methods. The reference would keep track of the
context information. This would allow sharing of graphs while maintaining context
traversability.
- Add Node[] findPaths(Node from, Node to, int maxLength, int maxPaths);
- What about simplify(RelationType, Concept[]) and simplify(Concept[])?
- Need some sort of "cutting" process that allows one to quickly obtain a subgraph.
However, see @idea tag in join() regarding excludeUnjoinedNodes flag.
Method Comments
- Method getActorRelations()
- Possibly cache results of this operation?
- Method getNormalRelations()
- Possibly cache results of this operation?
- Method join(notio.Graph, notio.Concept[], notio.Graph, notio.Concept[], notio.MatchingScheme, notio.CopyingScheme)
- Add exludeUnjoinedNodes flag. This means that only joined nodes will form the
final graph. This provides a simple way to obtain the graph for a projection by
joining two graphs on their matched nodes and excluding the rest. This is somewhat
complicated since joins are specified using concepts only. We need to include
information about mapped relations. Perhaps this is a distinct operation (e.g.
intersection).
Class Comments
- Change array constructor with starting output index to a more general
form that has two arrays and two indicies for each array.
Class Comments
- Should we allow people to query an implementation name here?
Class Comments
- RN_MATCH_QUANTIFIER?
- Unordered arc matching?
- Possibly should have RN_/CN_MATCH_OFF in order to specify that those nodes
should not be considered during matching. Only makes sense with CONN_MATCH_OFF.
- QF_MATCH_QUANTITY - referents must refer to the same quantity (e.g. 1 == 1,
3 == 3, every == every).
- QF_MATCH_SUBSET - first referent must match some subset of the second referent.
Thus, if the first quantifier is larger than the second, they will never be matched
by this flag. Unclear how this should be dealt with two universal quantifiers.
Class Comments
- Consider adding method for adding concepts 'out of order' with respect to
dominant/subordinate issues.
- Should probably remove defining label stuff from this class in the name of
"purity". Strictly speaking, defining labels are not part of CG's (the abstract
syntax) and are purely a feature of notations (e.g. CGIF or LF).
Class Comments
- Include proper sub/supertype matching?
Class Comments
- NameMatchers like MarkerMatchers?
Class Comments
- Consider extending "pop" operations to the label generator.
Class Comments
- Should some or all macros be handled by a preprocessor? It could
certainly make things like "If-Then" and universal quantification
macros much simpler.
- Option to trim trailing/leading spaces from comments.
- Make some Vectors global to improve speed.
Method Comments
- Method GraphComment()
- Should this be split so we have a non-terminal for javadocs?
Class Comments
- Improve formatting by allowing newlines/indentation before/after each node.
Class Comments
- Should we expose this class in the API? Might be generally useful.
8 warnings