Just an update for the wider developer community (and users, if they’re interested) about what’s coming next.
We are expecting to release Psi4 1.1 before the end of 2016. Anticipated features include F-SAPT, open-shell SAPT, improvements to the density-fitted MCSCF code, analytic DF-RHF Hessians, and numerous infrastructure improvements (mainly affecting developers) including an even newer CMake build system (with faster build times and improvements for external projects), upgrade from Boost Python bindings to pybind11 (and general removal of Boost), and “inversion” of Psi4 to make it directly callable from Python. We also hope to have the ability to chose among multiple integrals libraries (beyond just libint and liberd).
We will submit a paper on Psi4 1.1 to alert the wider community to what has changed since our previous publication on (alpha) Psi4. The paper will focus on software design decisions, advanced libraries, and interoperability capabilities of Psi4.
We plan to spend about a week on Boost removal, sometime in August
We plan to spend a few days working on extensions to call any of multiple integral libraries, sometime after the Boost/CMake work is completed (Sept-Oct timeframe)
We plan to have a Psi4 Workshop in November, probably Nov 11-12 (and recurring annually the second Fri-Sat of November)
The Psi4 Workshop will be held in the CCQC at the University of Georgia on November 4 and 5. You’ll find a list of hotels near to the CCQC below. I have no idea about availability at each.
For hotels 2-5, which are located in downtown Athens, there is an option to take a campus bus (free) on Friday. Taking the campus bus cuts the journey to CCQC down to about a 5 minute walk to the bus stop and a 15 minute bus ride. The bus route that runs from downtown to the CCQC does not operate on weekends. I’ll send information on that closer to Workshop day.
Please forward this message to anyone who may be interested!
At this year’s Psi4 Workshop we will be introducing Lightning Talks. We want to know what people are doing with Psi4 either using it or developing with it. Excellent examples of Lighting Talks can be found on YouTube from the recent SciPy conference that Lori B and Daniel S attended.
The talks will be limited to 5 minutes, including setup, after which you will be driven off stage. Slides are encouraged, especially those that punctuate the talk, rather than anchor it. Some general advice can be found here on giving Lightning Talks. By limiting the amount of time each person presents, we should be able to have many people speak. I encourage everyone to consider giving a Lightning Talk.
Possible topics include:
What I’m developing for Psi4.
How Psi4 has confused me this year.
What I’m working on that could be generalized for Psi4.
What interface to this data structure I wish Psi4 had.
Why I cringe when told use Psi4.
Psi4 would be perfect if it did X and why.
Please let me know if you are interested in giving a Lightning Talk by email. There will also be a sign-up sheet on Friday of the workshop.
The history rewrite is complete, and the repo is shrunk by 91%. At this point please remove all GitHub forks and clones and refork/clone, respectively.