Drasil Team Meeting - Tuesday, Date, 1:30 pm, ITB/225

[smiths]

Agenda

1. Blockers for @jackwyand

• Moving chunks to drasil-metadata
  • Currently, all of the chunks that are added to the basis chunkdb are common chunks used in all of the case studies (i.e., not internal to Drasil)
  • Discuss how to continue with this project, I think keeping the chunks in drasil-data and then creating a new package drasil-metadb is what should be done
• Just so I don't forget, I was wondering why running make swhs_debug does not work while swhs_diff works. I ask because running debug seems to ensure that all the warnings are printed. Not being able to run it for SWHS makes me worried I will miss something.

2. Status of open PRs for @jackwyand

• SWHS: remove iMod output duplicate #4246: Ready for review
• SWHSNoPCM: Split concepts list to populate SymbolMap and ConceptMap with correct chunks #4245: Ready for review
• GamePhysics: Move defSymbols to TermMap to avoid duplicates #4243: Ready for review (@balacij may want to look since it relates to the problematic seeds issue)
• Remove moment duplicate #4231: Ready for re-review
• Move meta concepts to drasil-metadata #4208: Blocked by previously discussed blocker
• Create basis ChunkDB that all new ChunkDBs are built upon #4165: Should be closed once the changes have been made into separate PRs
  • Move basis to new package
  • Update imports and case study chunkdbs to work properly with the basis

3. Blockers for @sarrasoussia
   I've been working on automating the Table of Abbreviations and Acronyms (TAandA), as outlined in issue Automate creation of Table of Abbreviations and Acronyms #698. The goal was to remove the hacky use of a dummy ChunkDB in the System record and instead pass the list of abbreviations directly into the TAandA constructor inside SRSDecl. Here's what I’ve done so far:

I updated the RefTab constructor:
TAandA :: [IdeaDict] -> RefTab
I removed _usedinfodb from System, and deleted the related dummy usedDB in examples like SSP.

I modified the mkSubRef and mkRefSec logic to accept the new TAandA [IdeaDict] form.

To generate the table, I used:
let db = cdb ([] :: [QuantityDict]) ideas [] [] [] [] [] [] [] [] [] []
followed by:
tableAbbAccGen $ nub $ map fst $ Map.elems $ termTable db
However, I'm now completely blocked on this recurring type error:
Ambiguous type variable ‘c0’ arising from a use of ‘cdb’ prevents the constraint ‘(ConceptDomain c0)’ from being solved.
Even after explicitly type-annotating ideas as [IdeaDict], and adding [QuantityDict] to the first argument, the ambiguity persists. I also tried ScopedTypeVariables, and attempted to bring the correct types into scope — but the compiler still can't infer c0.

At this point, I’m unsure whether the issue stems from how IdeaDict is resolved (as a type synonym for CI), or if there's something deeper in how cdb’s polymorphism interacts with the rest of the module.

If you have any suggestions or insights on how to guide the type inference correctly, I’d really appreciate your help. This part has become a major bottleneck, and I’d like to move forward with the TAandA refactor.
5. Status of open PRs for @sarrasoussia
6. Status of other open PRs.Regrets: Jacques Carette (off at an all-day training session)
Virtual:

[smiths]

You can look at Dong's thesis for more information on the double pendulum example.

[smiths]

All agenda items were discussed.

@JacquesCarette, when you get a chance, can you look at the PRs from @jackwyand that are ready for review (#4246, #4245, #4231).

@sarrasoussia is going to clarify the changes she made for #4223 via responses to the comments from @JacquesCarette on the PR. @sarrasoussia will also update stable so that we can see how the vectors are displayed in the SRS.

@balacij provided guidance on #698.