AP/DJ AP: a generic technology DJ: a low-learning-curve implementation of AP

Law of Demeter Principle Each unit should only have limited knowledge about other units: only about units “closely” related to the current unit. “Each unit should only talk to its friends.” “Don’t talk to strangers.” Main Motivation: Control information overload. We can only keep a limited set of items in short-term memory.

Law of Demeter FRIENDS

Application to OO Unit = method –closely related = methods of class of this/self and other argument classes methods of immediate part classes (classes that are return types of methods of class of this/self ) In the following we talk about this application of the Law of Demeter Principle to OO

Rumbaugh and the Law of Demeter Quote: Avoid traversing multiple links or methods. A method should have limited knowledge of an object model. A method must be able to traverse links to obtain its neighbors and must be able to call operations on them, but it should not traverse a second link from the neighbor to a third class.

Law of Demeter (alternative formulation) A method should have limited knowledge of an object model. AP is a reaction to this view of the Law of Demeter

Agreement that LoD Good Idea How to follow LoD: good solutions exist but not widely known. Two approaches to following LoD: –OO approach –Adaptive approaches DJ APPC Demeter/Java

What if your friends are far away? You pay them to travel to you or you send an agent to them to collect the information you need. –Approximate Directions: You give them or your agent directions about what kind of information to collect but you don’t care about accidental details of the travel. –Detailed Directions: You give them or your agent detailed travel directions.

Adaptive Following LoD FRIENDS S A b C X a:From S to A b:From S to B c:From S via X to C a c

Traversal specifications create friends Class ClassGraph is an intermediate class between classes that need to communicate ClassGraph.fetch(Object o, TravSpec s)

P1 P2 P3 P1 P6 P2 P5 P3 P4 P1 P2 Adaptive Following of LoD: Key Idea

DJ An implementation of AP using only the DJ library and the Java Generic Library (JGL) All programs written in pure Java Intended as prototyping tool: makes heavy use of introspection in Java Integrates Generic Programming (a la STL) and Adaptive programming

Integration of Generic and Adaptive Programming A traversal specification turns an object graph into a container. Can invoke 50+ generic algorithms on those containers. Examples: add, delete, find, etc. What is gained: genericity not only with respect to data structure implementations but also with respect to class graph

Algorithm Classes Applying Applies a function to every element of a sequence Comparing Mismatches, performs equality tests, performs lexicographical comparisons Copying Copies a sequence Counting Counts unconditionally and conditionally Filling Fills a sequence with a single element Filtering Filters a sequence Finding Finds an object or an element that satisfies a predicate

Algorithm Classes Hashing Contains generic hashing algorithms Heap Makes, pushes, pops, and sorts a heap MinMax Finds the min and max of a sequence OrderSetOperations Contains Generic set operation algorithms Permuting Cycles through permutations of a sequence Printing Prints sequences and containers Removing Removes an object or element that satisfies a predicate

Algorithm Classes Replacing Replaces an object or element that satisfies a predicate Reversing Reverses a sequence Rotating Rotates a sequence SetOperations Union, intersection, difference, and inclusion Shuffling Shuffles a sequence Sorting Sorts a sequence Swapping Swaps elements or sequences Transforming Maps one sequence to another

Collections JGL includes 11 highly optimized data structures that satisfy most programmer needs. These data structures have been engineered for performance and ease of use, and their performance either meets or beats every other commercially available Java container library. In addition, all JGL containers work correctly in multithreaded situations.

Collection Interface add(Object) Add an object to myself. clear() Remove all of my objects. clone() Return a shallow copy of myself. elements() Return an Enumeration of the components in this container equals(Object) Return true if I'm equal to a specified object. finish() Return an iterator positioned immediately after my last item.

Collection Interface isEmpty() Return true if I contain no objects. maxSize() Return the maximum number of objects that I can contain. remove(Enumeration) Remove the element at a particular position. remove(Enumeration, Enumeration) Remove the elements in the specified range. size() Return the number of objects that I contain. start() Return an iterator positioned at my first item. toString() Return a string that describes me.

com.objectspace.jgl.algorithms. Filtering reject(Container, UnaryPredicate) Reject elements in a container. reject(InputIterator, InputIterator, UnaryPredicate) Reject elements in a range. select(Container, UnaryPredicate) Select elements in a container. select(InputIterator, InputIterator, UnaryPredicate) Select elements in a range. …

com.objectspace.jgl.algorithms. Permuting nextPermutation(BidirectionalIterator, BidirectionalIterator, BinaryPredicate) Arrange a sequence to become its next permutation. nextPermutation(Container, BinaryPredicate) Arrange a container to become its next permutation. prevPermutation(BidirectionalIterator, BidirectionalIterator, BinaryPredicate) Arrange a sequence to become its previous permutation. prevPermutation(Container, BinaryPredicate) Arrange a container to become its previous permutation.

com.objectspace.jgl.algorithms. Transforming collect(Container, UnaryFunction) Return a container that is the same class as the original and contains the result of applying the given unary function to each element in the original. transform(Container, Container, Container, BinaryFunction) Traverse two containers and add the results of invoking a BinaryFunction on corresponding elements to another container.

Sample DJ code // Find the user with the specified uid Container libUsers = new Container(library, "from Library to User"); User user = libUsers.find("uid", uid);

Methods provided by DJ On all objects: traverse, fetch, gather On containers: Java Generic Library algorithms (in progress) traverse is the important method; fetch and gather are special cases

Traverse method: excellent support for Visitor Pattern // class ClassGraph Object traverse(Object o, TravSpec s, Visitor v); traverse navigates through Object o following traversal specification s and executing the before and after methods in visitor v ClassGraph is computed using introspection

Guidelines IF you plan to use 1 sg with 1 o and several v THEN freeze(cg,sg,o)->ogs 1 sg with several (o or v) THEN freeze(cg,sg)->tg 1 o with several (sg or v) THEN freeze(cg,o)->og 1 tg with 1 o and several v THEN freeze(tg,o)->ogs 1 sg with 1 (o and v) THEN inline cg class graph sgstrategy graph tgtraversal graph oobject ogobject graph vvisitor or generic algorithm Abreviations

Lack of parameterized classes in Java makes DJ harder to use Consider the traversal: from A to B Let’s assume that in the class graph between A and B there is a Java collection class. The intent is: A = List(B) which we cannot express in Java. Instead we have: A = Vector(Object). Object : A | B. Let’s assume we also have a class X=B.

Lack of parameterized classes in Java makes DJ harder to use We have: A = Vector(Object). Object : A | B | X. X = B. If the vector contains an X object it will be traversed!!! A X Object B Vector *

A X Object B Vector * A XB * No X-object is allowed to be in A-object

Moral of the story If the Collection objects contain only the objects advertised in the nice class graph of the application the traversal done by DJ will be correct. However, if the Collection objects contain additional objects (like an X-object) they might be traversed accidentally.

Size of traversal graph DJ might create big traversal graphs when collection classes are involved. DJ will plan for all possibilities even though only a small subset will be realized during execution. To reduce the size of the traversal graph, you need to use bypassing. In the example: from A bypassing {A,X} to B.

DJ and collaborations Collaborations have a participant graph and are used with connectors. DJ allows us to write the behavior of collaborations such that the connectors only need to map roles to classes. Mapping of edges will be done by behavior. Traversal strategy graphs can be viewed as participant graphs.

Expressing collaborations in DJ ClassGraph cg=new ClassGraph(); TraversalGraph tg = TraversalGraph.compute(cg, new Strategy("from UMLDiagram via Abstract to Vertex")); CountInhRelsVisitor v = new CountInhRelsVisitor(); tg.traverse( cg, v);

Expressing collaborations in DJ public class CountInhRelsVisitor extends Visitor { public int iCount; public void start() { System.out.println("begin"); iCount = 0;} public void finish() { System.out.println("The total count = " + iCount); System.out.println("end"); } public void before(Vertex o) { iCount++; } …}

Traversal-based collaborations Where are the participants? Three of them. –from UMLDiagram via Abstract to Vertex Where is the participant graph? UMLDiagramAbstract Vertex * *

Traversal-based collaborations Where is the behavior of the collaboration? –It is split into three parts: The traversal graph. Has the meaning of traversing entire participant graph –from UMLDiagram via Abstract to Vertex The visitor and The gluing of traversal graph and visitor

Collaborations and DJ DJ allows us to conveniently express traversal-based collaborations. –The positive part (delete the bypassing) of a traversal strategy qualifies as participant graph. –The negative part of a traversal strategy can be considered a part of the connector.

Non-traversal based collaborations? Unless the participant graph consists only of one node, there is always some traversal in a collaboration. Otherwise it would not be a collaboration. Writing collaboration behavior in DJ simplifies connectors and makes behavior higher level.

More info DJ Home Page