Copyright 2022, NavAbility, All rights reserved.

Formalizing Project Level Requirements

• Avoid the Gaussian-only road: https://github.com/JuliaRobotics/DistributedFactorGraphs.jl/issues/535
• Avoid OO Polymorphic only patterns: https://github.com/JuliaRobotics/DistributedFactorGraphs.jl/issues/157#issuecomment-672813937
• Signature cross feeding / symmetry: https://github.com/JuliaRobotics/DistributedFactorGraphs.jl/issues/264#issue-552973135
  • Exceptions:
  • https://github.com/JuliaRobotics/IncrementalInference.jl/issues/1424#issuecomment-945157845
• Serialization and Marshaling: https://github.com/JuliaRobotics/DistributedFactorGraphs.jl/issues/590
• highly multithreaded and performant (TBD on which issue comments to link)
• SuiteSparse as conditional dependency only https://github.com/JuliaRobotics/IncrementalInference.jl/issues/793#issue-660607972
• Consolidate / Simplify factor evaluation (CalcFactor) https://github.com/JuliaRobotics/IncrementalInference.jl/issues/467
• CSM and ParametricCSM should be interchangeable https://github.com/JuliaRobotics/IncrementalInference.jl/issues/558#issuecomment-672817293
• First state machine tree init (single core) https://github.com/JuliaRobotics/IncrementalInference.jl/pull/273#issue-276427786
• Finalize state for CSM loggers https://github.com/JuliaRobotics/IncrementalInference.jl/issues/321#issue-456712286
• consolidate solver debug output in randomized per solve output folder https://github.com/JuliaRobotics/IncrementalInference.jl/issues/338#issue-462812054 # should be a sub-bullet here somewhere
• Documenting CSM and inference steps for debugging https://github.com/JuliaRobotics/IncrementalInference.jl/issues/443#issue-526256697
• Transition to LightDFG as default(now renamed Graphs.jl) https://github.com/JuliaRobotics/IncrementalInference.jl/issues/364#issuecomment-529899343
• Consolidate belief messages and message channels https://github.com/JuliaRobotics/IncrementalInference.jl/issues/459#issue-529964851
• solve multiple orphaned graphs https://github.com/JuliaRobotics/IncrementalInference.jl/issues/518#issue-545466369
• revamp common SLAM frontend processor https://github.com/JuliaRobotics/Caesar.jl/issues/527#issue-574164820
• new on Dead Reckon Tether https://github.com/JuliaRobotics/Caesar.jl/issues/433#issue-527790188
• some use-cases want autoinit to be a separate process https://github.com/JuliaRobotics/IncrementalInference.jl/issues/229 # should be sub-bullet under tree init
• ROS Integration https://github.com/JuliaRobotics/Caesar.jl/issues/227
  • How to provide non-Gaussian with vanilla ROS msgs https://github.com/JuliaRobotics/Caesar.jl/issues/395#issue-506069860
• Snap package install https://github.com/JuliaRobotics/Caesar.jl/issues/278#issue-419596576
• Compile Static .so libraries for multi-language or executable use https://github.com/JuliaRobotics/Caesar.jl/issues/210
• Managing MIME types for bigData storage https://github.com/JuliaRobotics/Caesar.jl/issues/274#issue-419242832
• Generic N-ary multihypothesis (bad-init nullhypo addition) https://github.com/JuliaRobotics/IncrementalInference.jl/issues/248#issue-428326229
  • How to use the multi-hypothesis feature https://github.com/JuliaRobotics/Caesar.jl/issues/141#issuecomment-405422791
• Standardize null hypothesis in addFactor! https://github.com/JuliaRobotics/IncrementalInference.jl/issues/237#issue-424046072
• Federated tree solve requires easily separated up and down solves https://github.com/JuliaRobotics/IncrementalInference.jl/issues/307
• Feature Request: Solver Progress https://github.com/JuliaRobotics/IncrementalInference.jl/issues/224#issue-411962043
• getBestPointVal(fg, :x13), when to use max, mean, or on manifold moment https://github.com/JuliaRobotics/IncrementalInference.jl/issues/214#issue-409531374
• Ability to give drawGraph a file name https://github.com/JuliaRobotics/IncrementalInference.jl/issues/118#issue-342868429
• Allow for more flexible node names https://github.com/JuliaRobotics/IncrementalInference.jl/issues/165#issue-365003669
• Integrate with Manifolds.jl https://github.com/JuliaRobotics/RoME.jl/issues/244
• allow partial factors (TBD which issue comment to use)
• https://github.com/JuliaRobotics/IncrementalInference.jl/issues/783#issue-659285005
• https://github.com/JuliaRobotics/RoME.jl/issues/62#issuecomment-464063851 # must find better format and placement
• RoME Running on ARM: https://github.com/JuliaRobotics/RoME.jl/issues/35
• DFGFactor is immutable
  • https://github.com/JuliaRobotics/DistributedFactorGraphs.jl/issues/510
• One API for residual factors: https://github.com/JuliaRobotics/IncrementalInference.jl/issues/467#issuecomment-770261263
• Factors default to .Z: https://github.com/JuliaRobotics/IncrementalInference.jl/pull/1439#issuecomment-961107369
• What Manifold data formats (coords / tangents / points): https://github.com/JuliaRobotics/RoME.jl/issues/465
  • Deconv requires <:AbstractRelative measurements to be tangents: https://github.com/JuliaRobotics/IncrementalInference.jl/pull/1380#issue-992398675
  • Factors return tangent at p: https://github.com/JuliaRobotics/IncrementalInference.jl/pull/1292#issue-936927602
  • Deferring a LazyManifoldsWrapper: https://github.com/JuliaRobotics/RoME.jl/issues/450#issuecomment-944452280
• Factor design for Manifolds.jl integration: https://github.com/JuliaRobotics/RoME.jl/issues/465
  • Partial factor manifold definition: https://github.com/JuliaRobotics/RoME.jl/pull/495/files#r695912751
  • Method for partial manifold definition not yet finalized, see https://github.com/JuliaRobotics/Caesar.jl#726
  • "wrong factor reductions": https://github.com/JuliaRobotics/IncrementalInference.jl/issues/1507#issuecomment-1081087381
• CalcFactor.cache design, in-place or in-line: https://github.com/JuliaRobotics/RoME.jl/pull/562#issuecomment-1076405636

Will cleanup and expand as time goes on.