RoadmapDiscussion

From 2018-09-07 Roadmap

Promised not (totally?) delivered

• 5.1
• Complete beta?
• Complete paper
• Begin refactor on constraints/sim fits
• Sasfit and scatter models integrated (sasfit in marketplace ...scatter?)
• Sart on use P or S array to do P*S … or I/P?
• Starton Oriented 1D cuts fitted
• Start Advanced “how to fit my data” tutorial
• Start on General OZ solver
• PRISM incorporation
• McSAS integrated
• Architecture manual complete
• Multi CPU/GPU support??
• Identify “custom workflows”
• Custom model editor features added like including compare.py integration
• Headless operation for realtime analysis started
• sascalc separation started

Future promises

To Mid 2021

• 5.2
• Complete realtime analysis/headless workflow
• Complete sascald separation
• Complete of OZ solver
• Finish 1D cut fits and accepting P or S as array inputs
• Start on refactoring of Gen Scatt calculator?
• Begin refactoring P(r)
• Refactor plotting interface
• Refactor sim/const workflow interface
• Start work on new workflow interfaces

Mid 2022

• 5.3
• Finish refactoring of Gen Scatt calculator?
• Finish P(r) refactor
• Finish advanced model tutorial
• Finish outstanding documentation tasks (is that even a thing????)
• Prioritize new documentation tasks
• Include more new workflow interfaces
• Web UI begin
• Community meeting at SAS2021

Mid 2023

• 5.4? Or 6?
• Add better SAXS support
• ASAXS?? (maybe not?)
• Start work to enable GUI to launch jobs on cluster backends
• Start including intelligent help (AI?)
• Web UI continued

DISCUSSION of April 14, 2020

Items currently being worked on

• SasView paper
• beta approximation
• model reparameterization
• constrained fitting writ large
• McSAS integration

Items of high importance to one or more active developers

• separate out sascalc (seems to be getting in the way of what many people want to do)
• P(R) and Gen Scatt (Chan Zuckerberg grant)
• More complex structure factor support including more structure factors, times when P and S may not be strictly separable, partial structure factor approaches etc. Important among other for Hierarchichal structures and for PRISM and RISM)- (Yun may get a funded project for this)
• Engage broader community - develop true user community involvment (started with SAS2018 and aborted CodeCampX -Hige priority for future sustainability and keen interest for Paul Butler)

Wish list

• PRISM integration (Yun and/or Tyler)
• Improve speed/efficiency of computations (Adaptive integration? Other approaches? multi carg GPU support?)
• Futher Oriented systems work to truly make routine (will need to engage community more)
• Magnetic system. (Proposed code camp focused on this topic but will need to truly engage that community)
• Void size distribution workflows/algorithms (An important unsuported area but what would be needed? would McSAS integration be sufficient? or also need a PRINSAS rewrite? or a simulation approach?)
• “Comparison” to MD and MC simulations
• Do we move into dynamics: DLS and NSE? Is it a step to far? May need better way to read generic (not I vs Q) data.
• Think about how to integrate more general Q resolution into fitting structure.