2021-11-10 –, Room I
Ray Donnelly liked to say that software collections were defined by "islands of compatibility" - sets of software where the API and ABI requirements line up. Each package ecosystem defines their island differently, and each approach has advantages and disadvantages. This talk will compare the approaches of operating system maintainers, the greater conda ecosystem, and the somewhat ad-hoc status quo of the R world, in the hopes of making implicit assumptions and consequences explicit.
Michael started out in science, using Python and C++ for driving electron microscope equipment and analyzing data. This was in the bad old days of mostly needing to compile your own software to use the latest and greatest packages. He took his skills for building packages to Continuum Analytics (now Anaconda), where he maintained conda-build, conda, and many conda recipes for several years. There he also steered Continuum towards opening up their recipes and contributing to Conda-Forge. Lately he works at RStudio, where he fumbles with Go, SQL and Javascript in efforts to facilitate managing R and Python together.