The article mentions domain-specific libraries, such as
libxc[45], libint[46], ELSI[47], SIRIUS[48], M-A-D-N-E-S-S[49], or TiledArray[50]
as a way for the field to deal with the increasing specialization of computer hardware following the slowdown of single-core performance scaling.
Taking this argument to its logical conclusion, these libraries are increasingly going to be the new frontier for much of the cutting-edge development.
At the same time, as pointed out to me by @hyllios at the Psi-k conference, even popular libraries like libxc still receive little attention/recognition outside circles of code developers.
While citation information on the libraries may still be sparse, perhaps one step in the right direction could be to expand the paragraph on these libraries in the next update of the review, and instead of just listing the libraries to give a rough scope of each of their purpose and features.
Furthermore, it may be interesting to track the use of these domain-specific libraries in a separate column of the https://atomistic.software table (maybe shown only when explicitly selected).
Finally, while there is still a lot of work to be done to adopt standard citation practices for the simulation codes themselves, it may be useful to already include pointers to ideas/initiatives around transitive credit systems like this one that will be needed to properly credit the contribution of these libraries to the scientific process.
The article mentions domain-specific libraries, such as
as a way for the field to deal with the increasing specialization of computer hardware following the slowdown of single-core performance scaling.
Taking this argument to its logical conclusion, these libraries are increasingly going to be the new frontier for much of the cutting-edge development.
At the same time, as pointed out to me by @hyllios at the Psi-k conference, even popular libraries like libxc still receive little attention/recognition outside circles of code developers.
While citation information on the libraries may still be sparse, perhaps one step in the right direction could be to expand the paragraph on these libraries in the next update of the review, and instead of just listing the libraries to give a rough scope of each of their purpose and features.
Furthermore, it may be interesting to track the use of these domain-specific libraries in a separate column of the https://atomistic.software table (maybe shown only when explicitly selected).
Finally, while there is still a lot of work to be done to adopt standard citation practices for the simulation codes themselves, it may be useful to already include pointers to ideas/initiatives around transitive credit systems like this one that will be needed to properly credit the contribution of these libraries to the scientific process.