Open PLatCIC opened 4 months ago
A few observations:
So I'm not sure all of the problems you listed are the fault of VS. I think "trashed" is a bit over the top. I think it 10 minutes, you can delete those that you don't want and move on. I highly recommend using the datamap throughout the creation of a project instead of writing lots of rules and then doing the datamap. It is much more helpful as you are writing code.
I have a Visual Soar project with a few operators and some complex data structures. I worked hard to build the datamap by hand. Then I thought the automatic DM generation might find some things I missed, so I tried the Generate the Datamap from the Current Operator Hierarchy button. It made a long list of changes. Then when I looked with Display Top-State Datamap, I saw that it had severely trashed the datamap I had so carefully built by hand.
For example, on the top state there is an identifier ships and under it one called ship. Under ship there are a bunch of integer and string attributes. I had copied as a link the ships structure under several of the operators. It was still there as before, but now there is a new ship identifier under several of them that has all those attributes turned into identifiers. Similarly, there is a qna-query structure under io.output-link that was copied as a link under some of the operators. Now its sub-identifier query is copied on to the top level of the operator with its parts messed up.
To help document this, I'm attaching three files. ships-qna.zip has the whole project as it was before doing the automatic datamap generation. Then there are two versions of the .dm file, the one with -1 before the problem and the one with -2 after the problem. Both have .txt added to their names so that GitHub will accept them. It is obvious that the differences are enormous.
ships-qna-1.dm.txt ships-qna-2.dm.txt
It would be nice to have Visual Soar's help with building a datamap, but until this is fixed I will avoid using those features.