# Return to Question

 Notice removed Canonical answer required by Community♦ occurred Nov 29 '13 at 4:41 Bounty Ended with no winning answer by Community♦ occurred Nov 29 '13 at 4:41 6 Added another way to look at the problem edited Nov 22 '13 at 4:46 Geo 33922 silver badges1111 bronze badges Another way to look at it: The very concept of working with mutable trees violates the Law of Demeter. It wouldn't really be a violation of the law if the tree were a value as parse trees and syntax trees normally are, but in that case there would be no problem since nothing would need to be kept up-to-date. So then this problem exists as a direct result of violating the Law of Demeter, but how do you avoid that in general when your domain seems to be about manipulating trees or graphs? The Composite pattern is a wonderful tool for turning a graph into a single object and obeying the Law of Demeter. Is it possible to use the Composite pattern to effectively turn one kind of tree into another? Can you make a Composite parse tree so that it acts like an abstract syntax tree and even a control flow graph? Is there a way to do it without violating the single responsibility principle? The Composite pattern tends to cause classes to absorb every responsibility that they touch, but perhaps it could be combined with the Strategy pattern somehow. Another way to look at it: The very concept of working with mutable trees violates the Law of Demeter. It wouldn't really be a violation of the law if the tree were a value as parse trees and syntax trees normally are, but in that case there would be no problem since nothing would need to be kept up-to-date. So then this problem exists as a direct result of violating the Law of Demeter, but how do you avoid that in general when your domain seems to be about manipulating trees or graphs? The Composite pattern is a wonderful tool for turning a graph into a single object and obeying the Law of Demeter. Is it possible to use the Composite pattern to effectively turn one kind of tree into another? Can you make a Composite parse tree so that it acts like an abstract syntax tree and even a control flow graph? Is there a way to do it without violating the single responsibility principle? The Composite pattern tends to cause classes to absorb every responsibility that they touch, but perhaps it could be combined with the Strategy pattern somehow. Tweeted twitter.com/#!/StackProgrammer/status/403372661545238528 occurred Nov 21 '13 at 4:01 5 added 923 characters in body edited Nov 21 '13 at 3:18 Geo 33922 silver badges1111 bronze badges stylSupposeSuppose you have a parse tree, an abstract syntax tree, and a control flow graph, each one logically derived from the one before. In principle it is easy to construct each graph given the parse tree, but how can we manage the complexity of updating the graphs when the parse tree is modified? We know exactly how the tree has been modified, but how can the change be propagated to the other trees in a way that doesn't become difficult to manage? Edit: Let me elaborate on the troubles that this problem causes. This issue pops up in various places, whenever two parts of a project operate on graphs, with each graph being a different representation of the same thing that changes while the software is running. It's like making an adapter for an interface, but instead of wrapping a single object or a fixed number of objects, we need to wrap an entire graph of arbitrary size. Naturally we can't just use an ordinary graph conversion algorithm directly because that can't respond to changes in any way other than starting from scratch, so what are the alternatives? Perhaps the algorithm could be written in a continuation-passing style where each step of the algorithm is represented as an object with a method for each type of node in the original graph, like a visitor. Then the algorithm can be assembled by composing various simple visitors together. Another example: Suppose you have a GUI that is laid out like you might in Java Swing, using JPanels and layout managers. You can simplify that process by using nested JPanels in place of complex layout managers, so you end up with a tree of various containers that includes nodes that exist only for layout purposes and are otherwise meaningless. Now suppose that the same tree that is used to generate your GUI is also used in another part of your application, but instead of laying the tree out graphically it is working with a library that will generate an abstract representation tree as a system of folders. In order to use this library, we need to have a version of the tree that doesn't have the layout nodes; the layout nodes need to be flattened into their parent nodes, but the library still needs to be notified each time the tree changes as though the two versions of the tree were a single data structure. stylSuppose you have a parse tree, an abstract syntax tree, and a control flow graph, each one logically derived from the one before. In principle it is easy to construct each graph given the parse tree, but how can we manage the complexity of updating the graphs when the parse tree is modified? We know exactly how the tree has been modified, but how can the change be propagated to the other trees in a way that doesn't become difficult to manage?Edit: Let me elaborate on the troubles that this problem causes. This issue pops up in various places, whenever two parts of a project operate on graphs, with each graph being a different representation of the same thing that changes while the software is running. It's like making an adapter for an interface, but instead of wrapping a single object or a fixed number of objects, we need to wrap an entire graph of arbitrary size. Naturally we can't just use an ordinary graph conversion algorithm directly because that can't respond to changes in any way other than starting from scratch, so what are the alternatives? Perhaps the algorithm could be written in a continuation-passing style where each step of the algorithm is represented as an object with a method for each type of node in the original graph, like a visitor. Then the algorithm can be assembled by composing various simple visitors together. Suppose you have a parse tree, an abstract syntax tree, and a control flow graph, each one logically derived from the one before. In principle it is easy to construct each graph given the parse tree, but how can we manage the complexity of updating the graphs when the parse tree is modified? We know exactly how the tree has been modified, but how can the change be propagated to the other trees in a way that doesn't become difficult to manage?  Let me elaborate on the troubles that this problem causes. This issue pops up in various places, whenever two parts of a project operate on graphs, with each graph being a different representation of the same thing that changes while the software is running. It's like making an adapter for an interface, but instead of wrapping a single object or a fixed number of objects, we need to wrap an entire graph of arbitrary size. Naturally we can't just use an ordinary graph conversion algorithm directly because that can't respond to changes in any way other than starting from scratch, so what are the alternatives? Perhaps the algorithm could be written in a continuation-passing style where each step of the algorithm is represented as an object with a method for each type of node in the original graph, like a visitor. Then the algorithm can be assembled by composing various simple visitors together. Another example: Suppose you have a GUI that is laid out like you might in Java Swing, using JPanels and layout managers. You can simplify that process by using nested JPanels in place of complex layout managers, so you end up with a tree of various containers that includes nodes that exist only for layout purposes and are otherwise meaningless. Now suppose that the same tree that is used to generate your GUI is also used in another part of your application, but instead of laying the tree out graphically it is working with a library that will generate an abstract representation tree as a system of folders. In order to use this library, we need to have a version of the tree that doesn't have the layout nodes; the layout nodes need to be flattened into their parent nodes, but the library still needs to be notified each time the tree changes as though the two versions of the tree were a single data structure. Notice added Canonical answer required by Geo occurred Nov 21 '13 at 2:47 Bounty Started worth 50 reputation by Geo occurred Nov 21 '13 at 2:47 4 word choice edited Nov 21 '13 at 2:45 Geo 33922 silver badges1111 bronze badges SupposestylSuppose you have a parse tree, an abstract syntax tree, and a control flow graph, each one logically derived from the one before. In principle it is easy to construct each graph given the parse tree, but how can we manage the complexity of updating the graphs when the parse tree is modified? We know exactly how the tree has been modified, but how can the change be propagated to the other trees in a way that doesn't become difficult to manage? Naturally the dependent graph can be updated by simply reconstructing it from scratch every time the first graph changes, but then there would be no way of knowing the details of the changes in the dependent graph. I currently have four ways to attempt to solve this problem, but each one has difficulties. Nodes of the dependent tree each observe the relevant nodes of the original tree, updating themselves and the observer lists of original tree nodes as necessary. The conceptual complexity of this can become daunting. Each node of the original tree has a list of the dependent tree nodes that specifically depend upon it, and when the node changes it sets a flag on the dependent nodes to mark them as dirty, including the parents of the dependent nodes all the way down to the root. After each change we run an algorithm that is much like the algorithm for constructing the dependent graph from scratch, but it skips over any clean node and reconstructs each dirty node, keeping track of whether the reconstructed node is actually different from the dirty node. This can also get tricky. We can represent the logical connection between the original graph and the dependent graph as a data structure, like a list of constraints, perhaps designed using a declarative language. When the original graph changes we need only scan the list to discover which constraints are violated and how the dependent tree needs to change to correct the violation, all encoded as data. We can reconstruct the dependent graph from scratch as though there were no existing dependent graph, and then compare the existing graph and the new graph to discover how it has changed. I'm sure this is the easiest way because I know there are algorithms available for detecting differences, but they are all quite computationally expensive and in principle it seems unnecessary so I'm deliberately avoiding this option. What is the right way to deal with these sorts of problems? Surely there must be a design pattern that makes this whole thing almost easy. It would be nice to have a good solution for every problem of this general description. Does this class of problem have a name? Edit: Let me elaborate on the troubles that this problem causes. This issue pops up in various places, whenever two parts of a project operate on graphs, with each graph being a different representation of the same thing that changes while the software is running. It's like making an adapter for an interface, but instead of wrapping a single object or a fixed number of objects, we need to wrap an entire graph of arbitrary size. Every time I try this I end up with a confusing unmaintainable mess. The control flow of observers can be difficult to follow when it gets complicated, and the algorithm for converting one graph to another is usually tricky enough to follow when its laid out plainly and not spread across multiple classes. The problem is that there seems to be no way to use just a plain, straight-forward graph conversion algorithm when the original graph is changing. Naturally we can't just use an ordinary graph conversion algorithm directly because that can't respond to changes in any way other than starting from scratch, so what are the alternatives? Perhaps the algorithm could be written in a continuation-passing waystyle where each step of the algorithm is represented as an object with a method for each type of node in the original graph, like a visitor. Then the algorithm can be assembled by composing various simple visitors together. Suppose you have a parse tree, an abstract syntax tree, and a control flow graph, each one logically derived from the one before. In principle it is easy to construct each graph given the parse tree, but how can we manage the complexity of updating the graphs when the parse tree is modified? We know exactly how the tree has been modified, but how can the change be propagated to the other trees in a way that doesn't become difficult to manage? Naturally the dependent graph can be updated by simply reconstructing it from scratch every time the first graph changes, but then there would be no way of knowing the details of the changes in the dependent graph. I currently have four ways to attempt to solve this problem, but each one has difficulties. Nodes of the dependent tree each observe the relevant nodes of the original tree, updating themselves and the observer lists of original tree nodes as necessary. The conceptual complexity of this can become daunting. Each node of the original tree has a list of the dependent tree nodes that specifically depend upon it, and when the node changes it sets a flag on the dependent nodes to mark them as dirty, including the parents of the dependent nodes all the way down to the root. After each change we run an algorithm that is much like the algorithm for constructing the dependent graph from scratch, but it skips over any clean node and reconstructs each dirty node, keeping track of whether the reconstructed node is actually different from the dirty node. This can also get tricky. We can represent the logical connection between the original graph and the dependent graph as a data structure, like a list of constraints, perhaps designed using a declarative language. When the original graph changes we need only scan the list to discover which constraints are violated and how the dependent tree needs to change to correct the violation, all encoded as data. We can reconstruct the dependent graph from scratch as though there were no existing dependent graph, and then compare the existing graph and the new graph to discover how it has changed. I'm sure this is the easiest way because I know there are algorithms available for detecting differences, but they are all quite computationally expensive and in principle it seems unnecessary so I'm deliberately avoiding this option. What is the right way to deal with these sorts of problems? Surely there must be a design pattern that makes this whole thing almost easy. It would be nice to have a good solution for every problem of this general description. Does this class of problem have a name? Edit: Let me elaborate on the troubles that this problem causes. This issue pops up in various places, whenever two parts of a project operate on graphs, with each graph being a different representation of the same thing that changes while the software is running. It's like making an adapter for an interface, but instead of wrapping a single object or a fixed number of objects, we need to wrap an entire graph of arbitrary size. Every time I try this I end up with a confusing unmaintainable mess. The control flow of observers can be difficult to follow when it gets complicated, and the algorithm for converting one graph to another is usually tricky enough to follow when its laid out plainly and not spread across multiple classes. The problem is that there seems to be no way to use just a plain, straight-forward graph conversion algorithm when the original graph is changing. Naturally we can't just use an ordinary graph conversion algorithm directly because that can't respond to changes in any way other than starting from scratch, so what are the alternatives? Perhaps the algorithm could be written in a continuation-passing way where each step of the algorithm is represented as an object with a method for each type of node in the original graph, like a visitor. Then the algorithm can be assembled by composing various simple visitors together. stylSuppose you have a parse tree, an abstract syntax tree, and a control flow graph, each one logically derived from the one before. In principle it is easy to construct each graph given the parse tree, but how can we manage the complexity of updating the graphs when the parse tree is modified? We know exactly how the tree has been modified, but how can the change be propagated to the other trees in a way that doesn't become difficult to manage? Naturally the dependent graph can be updated by simply reconstructing it from scratch every time the first graph changes, but then there would be no way of knowing the details of the changes in the dependent graph. I currently have four ways to attempt to solve this problem, but each one has difficulties. Nodes of the dependent tree each observe the relevant nodes of the original tree, updating themselves and the observer lists of original tree nodes as necessary. The conceptual complexity of this can become daunting. Each node of the original tree has a list of the dependent tree nodes that specifically depend upon it, and when the node changes it sets a flag on the dependent nodes to mark them as dirty, including the parents of the dependent nodes all the way down to the root. After each change we run an algorithm that is much like the algorithm for constructing the dependent graph from scratch, but it skips over any clean node and reconstructs each dirty node, keeping track of whether the reconstructed node is actually different from the dirty node. This can also get tricky. We can represent the logical connection between the original graph and the dependent graph as a data structure, like a list of constraints, perhaps designed using a declarative language. When the original graph changes we need only scan the list to discover which constraints are violated and how the dependent tree needs to change to correct the violation, all encoded as data. We can reconstruct the dependent graph from scratch as though there were no existing dependent graph, and then compare the existing graph and the new graph to discover how it has changed. I'm sure this is the easiest way because I know there are algorithms available for detecting differences, but they are all quite computationally expensive and in principle it seems unnecessary so I'm deliberately avoiding this option. What is the right way to deal with these sorts of problems? Surely there must be a design pattern that makes this whole thing almost easy. It would be nice to have a good solution for every problem of this general description. Does this class of problem have a name? Edit: Let me elaborate on the troubles that this problem causes. This issue pops up in various places, whenever two parts of a project operate on graphs, with each graph being a different representation of the same thing that changes while the software is running. It's like making an adapter for an interface, but instead of wrapping a single object or a fixed number of objects, we need to wrap an entire graph of arbitrary size. Every time I try this I end up with a confusing unmaintainable mess. The control flow of observers can be difficult to follow when it gets complicated, and the algorithm for converting one graph to another is usually tricky enough to follow when its laid out plainly and not spread across multiple classes. The problem is that there seems to be no way to use just a plain, straight-forward graph conversion algorithm when the original graph is changing. Naturally we can't just use an ordinary graph conversion algorithm directly because that can't respond to changes in any way other than starting from scratch, so what are the alternatives? Perhaps the algorithm could be written in a continuation-passing style where each step of the algorithm is represented as an object with a method for each type of node in the original graph, like a visitor. Then the algorithm can be assembled by composing various simple visitors together. 3 Added additional explaination and idea edited Nov 13 '13 at 6:54 Geo 33922 silver badges1111 bronze badges 2 Replacing "listener" with "observer" edited Nov 9 '13 at 12:56 Geo 33922 silver badges1111 bronze badges 1 asked Nov 9 '13 at 10:27 Geo 33922 silver badges1111 bronze badges