First pass for generic DSM - Moose-dev

Stéphane Ducasse

21 Apr 21 Apr

1:38 p.m.

good! On Apr 21, 2010, at 12:38 PM, Laval Jannik wrote:

Hi all, I am trying to make DSM generic. In the last version of DSM, we can make a DSM based on nodes and edges. Only the basic DSM works on it (dsm with colors). And I have to debug the interaction of the visualization. In the package DSMCore, I made a class DSMExample with, for now, 2 examples. Here is an example: ==== dSMMatrix := DSMMatrix withNodes: #(#1 #2 #3) edges: #(#(#1 #2) #(#2 #1) #(#3 #1)). view := MOViewRenderer new. DSMVisualization new open: dSMMatrix on: view. view open ==== We make a DSM with nodes and a collection of edges. An edge is a collection of two elements: the source and the target. I think I will do a third element to add informations for other DSM. You can try it, and make comments. I will pay attention to your idea to have a better generic DSM. Cheers --- Jannik Laval _______________________________________________ Moose-dev mailing list Moose-dev(a)iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev

Reply

Simon Denier

4:17 p.m.

On 21 avr. 2010, at 15:41, Alexandre Bergel wrote:

...

Cool! It works for me.

dSMMatrix := DSMMatrix withNodes: #(#1 #2 #3) edges: #(#(#1 #2) #(#2 #1) #(#3 #1)).

I was wondering why not to use the Mondrian convention for specifying edges. In mondrian you can specify edges with: view edges: nodes from: #selector1 to: #selector2.

Indeed, this API is also reused by MooseAlgos Graph. Maybe we could define a trait.

...

Also, maybe it would be nice to have a DSMShape in Mondrian, that draw a DSM for a particular node (e.g., a group of packages). But maybe this could be quite some work to adapt your implementation.

I am not sure I understand what you mean. The generic DSM could/should be customizable in three ways: - input model - a DSMCellShape, which takes as input the row, the column, and the dsm model - interactions (but this comes with the CellShape normally, except for a few global interactions) -- Simon

Reply

Simon Denier

4:50 p.m.

On 21 avr. 2010, at 16:19, Tudor Girba wrote:

...

Hi, That is great indeed. Jannik, thanks for spending the time to make it generic. It will be very useful. I agree with the suggestion of Alex regarding the API. Perhaps you can build a DSMBuilder that just provides the API to populate your internal model.

While we are on it, I just want to put forward that the Graph library already comes with a generic graph builder just for this purpose (it's used for example in layer table and other places, to build graph structure with more or less arbitrary classes). The following is the sample expression to create a basic graph. MOGraphStructure new nodeClass: MOGraphNode; edgeClass: MOGraphEdge; nodes: #(1 2 3); edges: #((1 2) (2 1) (3 1)) from: #first to: #second.

...

Cheers, Doru On 21 Apr 2010, at 15:41, Alexandre Bergel wrote:

Cool! It works for me.

dSMMatrix := DSMMatrix withNodes: #(#1 #2 #3) edges: #(#(#1 #2) #(#2 #1) #(#3 #1)).

I was wondering why not to use the Mondrian convention for specifying edges. In mondrian you can specify edges with: view edges: nodes from: #selector1 to: #selector2. For each node contained in nodes, it creates an edge going from 'node selector1' to 'node selector2'. This is quite convenient, because I do not have to specify how each edge is defined within the script. The domain I wish to represent is in charge of defining the edges. If you still wish to explicitly defining edges, you can always do: -=-=-=-=-=-=-=-=-=-=-=-=-=-=-= view shape rectangle size: 30; withText. view nodes: (1 to: 3). view shape arrowedLine. view edges: {1 -> 2. 2 -> 1 . 3 ->1} from: #key to: #value. view treeLayout -=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Also, maybe it would be nice to have a DSMShape in Mondrian, that draw a DSM for a particular node (e.g., a group of packages). But maybe this could be quite some work to adapt your implementation. Cheers, Alexandre -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;. _______________________________________________ Moose-dev mailing list Moose-dev(a)iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev

-- www.tudorgirba.com "What is more important: To be happy, or to make happy?" _______________________________________________ Moose-dev mailing list Moose-dev(a)iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev

-- Simon

Reply

Laval Jannik

24 Apr 24 Apr

2:47 p.m.

Hi, I made it. Now, there are two ways to build a DSM. First, with the DSMBuilder, seond by a small script. Example of the first method: ===== DSMBuilder buildWithNodes: #(1 2 3 4 5 6) edges: #((1 2) (2 1) (3 1) (2 3) (4 5) (5 4) (6 4)) from: #first to: #second ===== Example of the second one: ===== |DSMAlgo dSMMatrix view| view := MOViewRenderer new. dSMMatrix := DSMMatrix withNodes: #(1 2 3 4 5 6) edges: #((1 2) (2 1) (3 1) (2 3) (4 5) (5 4) (6 4)) from:#first to:#second. DSMVisualization new open: dSMMatrix on: view. view open ===== Cheers, Jannik On Apr 21, 2010, at 16:50 , Simon Denier wrote:

...

On 21 avr. 2010, at 16:19, Tudor Girba wrote:

Hi, That is great indeed. Jannik, thanks for spending the time to make it generic. It will be very useful. I agree with the suggestion of Alex regarding the API. Perhaps you can build a DSMBuilder that just provides the API to populate your internal model.

While we are on it, I just want to put forward that the Graph library already comes with a generic graph builder just for this purpose (it's used for example in layer table and other places, to build graph structure with more or less arbitrary classes). The following is the sample expression to create a basic graph. MOGraphStructure new nodeClass: MOGraphNode; edgeClass: MOGraphEdge; nodes: #(1 2 3); edges: #((1 2) (2 1) (3 1)) from: #first to: #second.

Cheers, Doru On 21 Apr 2010, at 15:41, Alexandre Bergel wrote:

Cool! It works for me.

dSMMatrix := DSMMatrix withNodes: #(#1 #2 #3) edges: #(#(#1 #2) #(#2 #1) #(#3 #1)).

I was wondering why not to use the Mondrian convention for specifying edges. In mondrian you can specify edges with: view edges: nodes from: #selector1 to: #selector2. For each node contained in nodes, it creates an edge going from 'node selector1' to 'node selector2'. This is quite convenient, because I do not have to specify how each edge is defined within the script. The domain I wish to represent is in charge of defining the edges. If you still wish to explicitly defining edges, you can always do: -=-=-=-=-=-=-=-=-=-=-=-=-=-=-= view shape rectangle size: 30; withText. view nodes: (1 to: 3). view shape arrowedLine. view edges: {1 -> 2. 2 -> 1 . 3 ->1} from: #key to: #value. view treeLayout -=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Also, maybe it would be nice to have a DSMShape in Mondrian, that draw a DSM for a particular node (e.g., a group of packages). But maybe this could be quite some work to adapt your implementation. Cheers, Alexandre -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;. _______________________________________________ Moose-dev mailing list Moose-dev(a)iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev

-- www.tudorgirba.com "What is more important: To be happy, or to make happy?" _______________________________________________ Moose-dev mailing list Moose-dev(a)iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev

-- Simon _______________________________________________ Moose-dev mailing list Moose-dev(a)iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev

--- Jannik Laval

Reply

Alexandre Bergel

10:22 p.m.

Excellent! Alexandre On 24 Apr 2010, at 08:47, Laval Jannik wrote:

...

Hi, I made it. Now, there are two ways to build a DSM. First, with the DSMBuilder, seond by a small script. Example of the first method: ===== DSMBuilder buildWithNodes: #(1 2 3 4 5 6) edges: #((1 2) (2 1) (3 1) (2 3) (4 5) (5 4) (6 4)) from: #first to: #second ===== Example of the second one: ===== |DSMAlgo dSMMatrix view| view := MOViewRenderer new. dSMMatrix := DSMMatrix withNodes: #(1 2 3 4 5 6) edges: #((1 2) (2 1) (3 1) (2 3) (4 5) (5 4) (6 4)) from:#first to:#second. DSMVisualization new open: dSMMatrix on: view. view open ===== Cheers, Jannik On Apr 21, 2010, at 16:50 , Simon Denier wrote:

On 21 avr. 2010, at 16:19, Tudor Girba wrote:

Hi, That is great indeed. Jannik, thanks for spending the time to make it generic. It will be very useful. I agree with the suggestion of Alex regarding the API. Perhaps you can build a DSMBuilder that just provides the API to populate your internal model.

While we are on it, I just want to put forward that the Graph library already comes with a generic graph builder just for this purpose (it's used for example in layer table and other places, to build graph structure with more or less arbitrary classes). The following is the sample expression to create a basic graph. MOGraphStructure new nodeClass: MOGraphNode; edgeClass: MOGraphEdge; nodes: #(1 2 3); edges: #((1 2) (2 1) (3 1)) from: #first to: #second.

Cheers, Doru On 21 Apr 2010, at 15:41, Alexandre Bergel wrote:

Cool! It works for me. > dSMMatrix := DSMMatrix withNodes: #(#1 #2 #3) edges: #(#(#1 #2) > #(#2 #1) #(#3 #1)). I was wondering why not to use the Mondrian convention for specifying edges. In mondrian you can specify edges with: view edges: nodes from: #selector1 to: #selector2. For each node contained in nodes, it creates an edge going from 'node selector1' to 'node selector2'. This is quite convenient, because I do not have to specify how each edge is defined within the script. The domain I wish to represent is in charge of defining the edges. If you still wish to explicitly defining edges, you can always do: -=-=-=-=-=-=-=-=-=-=-=-=-=-=-= view shape rectangle size: 30; withText. view nodes: (1 to: 3). view shape arrowedLine. view edges: {1 -> 2. 2 -> 1 . 3 ->1} from: #key to: #value. view treeLayout -=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Also, maybe it would be nice to have a DSMShape in Mondrian, that draw a DSM for a particular node (e.g., a group of packages). But maybe this could be quite some work to adapt your implementation. Cheers, Alexandre -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;. _______________________________________________ Moose-dev mailing list Moose-dev(a)iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev

-- www.tudorgirba.com "What is more important: To be happy, or to make happy?" _______________________________________________ Moose-dev mailing list Moose-dev(a)iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev

-- Simon _______________________________________________ Moose-dev mailing list Moose-dev(a)iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev

--- Jannik Laval _______________________________________________ Moose-dev mailing list Moose-dev(a)iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev

-- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.

Reply