DCM Notes for 2024-10-21

[glopesdev]

Conference

• Review common session schedule and structure
• Review action plan for registrations

[glopesdev]

Feedback on overall talk structure:

• Start with current state of the topic
• End intro talks with ideas for what is coming soon
  • Prime feedback and questions

Example (Custom package development / Docs):

• C# tips and tricks for custom package development
  • Collect a series of examples of interesting patterns to address common challenges:
    • Custom converters
    • Custom editors (image pickers, file pickers, custom drop-downs, etc)
    • Custom GUIs
    • Reusing patterns inside custom code
  • Consider whether it may be worth talking about or showcasing different manager / state design patterns:
    • Arduino vs Harp vs Shaders
    • Probably this is going in too deep, but just mention that different patterns do exist and ideally should be followed

Feedback on overall Custom package development structure:

• Start with current state of Bonsai documentation (DocFX)
• Documentation provenance
• Get feedback from what people think would be helpful to get started with new packages, with Bonsai, etc

Extra topics to discuss:

• Add discussion on package icons
• Highlight at the start of the conference that participation in these discussions will actively shape the future of the language
• Describe some of the potential pitfalls for package interoperability
• Describe the role of Bonsai Foundation in the package ecosystem:
  • Maintenance of packages, CLA, distribution channels, etc.

General observations for conference:

• Consider waitlist for registrations, feel free to register for virtual attendance
• Investigate whether SWC recording system can support streaming to any webinar platform or restricted

Hardware session:

• Overall topics (Harp, Ephys, Cameras)
• Are we missing important classes of hardware?
• For cameras:
  • Are we covering a sufficient amount of models / vendors, or are there still important omissions?
  • Should we aim to cover new models 1:1 as we have been doing by integrating each SDK or aim to cover standards such as Genicam which could potentially open up a much broader space of cameras?
  • 3D cameras:
    • messy since vendors and models obsolete very quickly
    • custom applications for reconstruction, SLAM, etc, which obsoletes even faster
    • usually end application and software stack are the problem (e.g. how to get skeletons, image reconstruction, recognition, etc)
    • low-hanging fruit is integrating RealSense and make a path to log depth video
    • depth resolution and working distance are often incompatible with setups
  • How to make the process of integrating new hardware sustainable? Approach vendors directly? Have a system of bounties? Support and polish community contributions?
• Is it worth approaching the topic of 2P imaging:
  • Current approaches are very monolithic (100s of knobs, though each system might use only 10 of them and still throw in a few custom scripts)
  • Includes lots of stage / motor control
  • Would it be interesting to get integration with existing systems?
  • How to deal with implementation bifurcations, e.g. galvo vs resonant scanners.