As a simple experiment or learning experience, why not start with the same languages as Light Table, then add R and Julia. (or start with Light Table and add a spreadsheet view)
But if you wanted to raise the bar on spreadsheets in general, you might back away from the scripting end and think about sets, series, logic, and ontologies. Look at CLIPS and Prolog, and try out Protégé.
I would personally enjoy being able to define named streams with Redis-like pub/sub semantics driven by arbitrary generators, composeable as sets (when finite, or as constrained by rules), and having a parametric type system available when it makes sense. Then scripting.
Comments
As a simple experiment or learning experience, why not start with the same languages as Light Table, then add R and Julia. (or start with Light Table and add a spreadsheet view)
But if you wanted to raise the bar on spreadsheets in general, you might back away from the scripting end and think about sets, series, logic, and ontologies. Look at CLIPS and Prolog, and try out Protégé.
Also, "Populous is a generic tool for building ontologies from simple spreadsheet like templates." (http://www.e-lico.eu/populous.html)
I would personally enjoy being able to define named streams with Redis-like pub/sub semantics driven by arbitrary generators, composeable as sets (when finite, or as constrained by rules), and having a parametric type system available when it makes sense. Then scripting.