[Interest] Interest Digest, Vol 25, Issue 32

[Wegner]

Hi mandeep, 

thank you for your quick comments and suggestions.

I was also thinking of sub states for tasks as each task will have its own logic. It will be activated by the overall application state machine. Great to hear, that this is a good way to go. And I really like the approach of parallel state machines because this exactly minimizes the amount of transition and thus the complexity (if used properly).

Storing all events to replay on crash startup is error prone, I agree. Especially as we will habe intense user interaction.

By "data" I meant that during runtime the user has to define a lot of parameters and those are stored in a block of data. As the goal of the application is to gather all information within that data block this also can be used for recover mode. The data specifies which steps still need to be done and thus transitions (guards) can be defined that analyse the data and automatically go all the way to the right state. I will check on this idea with a prototype implementation and see how intense the state machine pattern will get.

Best Regards, Ingmar

P.S. ..second try again because of this horrible HTML mail mess most of the apps produce...