Input/Output and Proper Workflow of ASAPPlanner

[Selvomega]

Currently it seems that ASAPPlanner optimizes with several steps, each with some input and output.

Regarding input / output, I am still confused about

• what are the inputs user can control
• what are necessary,
• what are optional,
• what is the compromise if I leave optional inputs as default value

Regarding workflow, I am confused about

• Among all steps I can take for optimization, are they all necessary or some of them are optional (e.g. Claude told me that I can get a plan without the lifecycle step, but with some compromise which I cannot quite understand :-) )
• What are all possible ways a user can use this planner properly (this means, if some steps are optional as mentioned above, then it means user can use this repo differently by selecting only the steps it need), and what each choice promises / does not promise. (That said, ideally this should not happen, since the cleanest abstraction should only have 1 input and 1 output interfaces. However according to Claude this seems to be true.)

Why Claude does not help me understand this code base. My reasonings are

• Some designs do not have clear document or only have stale document, such that they are inherently ambiguous. Claude can only infer author's intension from the code.
• Claude will always try giving a conclusion in the ambiguous scenario, even it itself also does not actually know the answer. Such conclusions seem correct at first but later will make me feel something's off.
• e.g. I asked Claude how I can use ASAPPlanner to get a DAG, and it told me how to do it without the lifecycle step. Therefore I didn't even realize there's a lifecycle stuff until this morning. After realizing this, I asked Claude if this lifecycle stuff is necessary or option, and it just gave me some gibberish about reasons to use/skip the lifecycle step. I tried hard but cannot understand either side. :-)