Real-time software citations that can load software into a VM and render a cited research data set in the browser.
The question is then what kind of future science scenarios open up? Where does this take us?
(in a not to distant future — as this model has been in use for a few decades, for example by packet management systems like Debian’s Dpkg, etc.)
A researcher finds a reference to a 'research data set' with an accompanying ‘software citation’.
Systems are advanced to a point where the researcher can load the research data from a 'persistent identifier' (PID) link in the browser, and the browser will locate and load the exact version of the cited software into a 'virtual machine' (VM) of some sorts, allowing full use of the 'research data set' in the browser.
The hypothetical workflow:
The tool chain – It can be imagined that something like @ProjectJupyter is used where data and code can be edited and run, with #binder https://mybinder.org/ or @CodeOceanHQ or @SWHeritage being extended to offer VM services to the software makers. The glue in the citations for the 'research data' and 'software citations', can be DOIs from @ZENODO_ORG that have been wrapped in a future version of the #CFF Citation File Format @stdruskat, or #codemeta https://github.com/codemeta @cboettig