PRO_CHANGE_MANAGER_R0

[Astea]

wiki page: PRO_CHANGE_MANAGER_R0 - effort: 0.5d

[deyan]

Adding category

[deyan]

correcting category

[orliin]

• In my intuitive understanding there are states and changes are transitions between states. For example states A, B and C, trnasition 1 between A and B, transition 2 between B and C: A <-1-> B <-2-> C

In these terms redoing 1 is to go from state A to state B; undoing 1 is to go from state B to state A; skipping B is to group the two transitions 1 and 2 into one big transition (A <-1+2-> C); skipping transitions or grouping states sounds unnatural.

• skip/unskip and undo/redo are all going to be defined as changes and a change can be skipped - this means you can skip a skip and then skip this skip and so on - this could get complicated :)

Having these in mind, the analysis is either messed up or needs to have more detailed explanations of terminology and actions.

Note 1: if skipping means grouping of changes, then the problem from the last row of the example will not exist because you can only group existing changes that are surrounded by existing states.
Note 2: Milo said UnmanagedChange is just a change that has a special undo and redo, but still has them. This may mean that it is not necessary to always skip such changes (if skipping a change at all exists)

[peko]

Yes the changes could be defined as transitions. This is reasonable, but this is a how thing not a what thing. How things are going to be put in design and not in analysis.

Skipping a skip should also be possible and why not. This is the only way to group several transitions:

1. a = 5 - change 1 | a = 5
2. a = 7 - change 2 | a = 7
3. undo - change 3 | a = 5
4. skip change 3 - change 4 | a = 7
5. skip change 4 - change 5 | a = 5 - which is exactly the undo in change 3.

the result of all these will be a value of a that is 5. And this is complicated but as Milo said it should be possible.

Re - Note 1: the things in the analysis is about what should be done. The last step is something that should be made clear in the design. There are situations where things should simply be defined, but this is a design thing.

Re - Note 2: The unmanaged change is fixed and added in note so that the person designing the task is aware of it.

[gogov]

• Maybe some more specific comments of what stuff is to be reviewed in the code of the UndoManager could be added
• It could be noted that skip and unskip accept an absolute command number as argument
• I'd like to note that the numbering of the commands in the examples provided in the Implementation Idea section doesn't change, because a person could get the impression that when skipping a given command, all following commands decrease their numbers with one

[deyan]

Batch update from file query-obsoleted.csv