| MT – JW – Interest in RDKit UGM? MT – Might not be able to swing logistics for this. This is in a patch of weddings for me. Also I might save my dry powder for more relevant workshops. MT – Not objecting to conference attendance in general, would like to figure out conferences where I can represent OpenFF/network effectively. MT – In future years this could be cool - https://fomms.cache.org/
I’m going to make an effort to spread out Interchange knowledge to me and JM. I’ve tasked him with trying out import/export and raising issues/making fixes. Also, if you can think of a high-level “theory” section that lays out your understanding/the theory of the package (maybe an extension of your already-nice “Capabilities” page to be more extensive? Or could consolidate knowledge here?) MT – Agree this is important, need to think about how this could be handled best. I’m not sure there’s an easy answer - Code readability can handle low-level stuff. Not sure how to allocate effort on this. Maybe we could take a few angles on this - some added documentation, some developer meetings, etc. JW – SB left the “how to train your force field” blog post which was amazing. So maybe blog post or live coding/stream of consciousness video for making a new importer/exporter. Could also spin out small PRs to JM and I, discussions and reviews will highlight misunderstandings. And possibly grand-strategy docs page, which could be uselessly abstract, but kinda encapsulate your high level thinking on what’s in scope and not, and how information maps into different items, and likely big challenges coming ahead. MT – Would add something FAQish between developer and user space. eg “how do I do X”, “how do I get all the bond force constants”, common things in importers/exporters. Some pushback on spinning out small PRs to you and JM, since that is supply-limited by the number of meaningful small PRs. Since we’re constrained on developer-time, a lot of PRs go un-reviewed. Some of the GROMACS performance stuff might have been caught by review (eg unnecessary lines in loops that could have been outside). So what I’m thinking is that it could be successful for me to do some documentation on how things work, and then you and JM could identify things that could be done better and submit PRs. JW – I think that’s solid. It’s likely we could jump in blind but more docs would be great. MT – So I think there are two things that could be done here:
Two weeks ago, I was tasked with doing a knowledge assessment/cross-training plan. Am I missing anything big here? MS is likely to land a grant for a software scientist working under his leadership, mentored by me to the tune of 0.5-1 month/year. Will work on polymer infrastructure build on top of OpenFF ecosystem. Funding may begin as early as September. MT – More details about this grant? CSSI is from NSF. MoSDeF had a few cycles of funding from this. PCummings ran this as a multi-lab collaboration. JW – I don’t think this is (JW looks at application PDF in #funding channel) JW – Oh, yeah, this IS a big multi-center thing. MT – Aiming to become a new OMSF project? JW – Unsure.
|