Closed source's future in Big Science clouds?
I follow a lot of astronomical/astrophysical missions and notice they have now gotten massive in size, scope, and data. Petabyte datasets are now commonplace. Some are now moving essentially to private clouds where researchers create accounts and use primarily FOSS tools to process data. Some examples are the Vera Rubin Observatory's Rubin Science Platform and ESA's Datalabs. Jupyter has entrenched itself deeply despite its many shortcomings. I can't create accounts on any of these platforms but I am guessing that getting support for license-servers and other paraphernalia associated with closed-source software won't be easy.
It makes sense at some level when datasets get so big they can't be practically served to a researcher's machine. But, I do wonder what this public data in private clouds means for MATLAB and similar paid software. All the projects say that FOSS encourages reproducibility but I've been burned more than once in my working life by randomly phased Python package updates and abandoned projects.
I'm retired now and don't have a dog in the fight. My MATLAB home license is just for self-improvement. I was just curious whether Mathworks and similar providers will have to give up on the basic science community and focus on applied, mission-critical areas where "some guy on Github" is not sufficient traceability.
Tags
