This reads as another argument about intentional systems versus emergent systems. The author wants to intentionally build a language which is provably "better". Instead we have an ecosystem where many languages (and libraries, and frameworks, and IDEs, etc.) are created, each with varying attributes, and of varying quality; some emerge as preferred for various applications, others fall by the wayside. And then the cycle begins anew. The author seems to miss the important lessons from the "market process" at work.
Perhaps languages are no longer at the right granularity to be "solved". This again echoes the market process; too many variables, and too many different preferences leads would-be problem solvers to rely on the pretense of knowledge.
Languages combine a host of disparate features, and sometimes simply combining previously known features in novel ways is sufficient iterative improvement. Clojure is one good example of this, where immutability, functional programming, and STM were all known and used to varying degrees, but combining them in a clever way allowed for something greater to emerge, particularly how the first two allowed for a new form of the third. From this in turn emerged a (new?) well thought out model of time state, and identity.
In the end, perhaps languages are now really engineering problems, not the science/math problems they used to be. Perhaps the academics should embrace this. To continue the economic parallel, if one wishes to examine the qualitative aspects of PLs, perhaps research should approach from the historical perspective, asking why certain languages emerged to solve certain problems.
Perhaps languages are no longer at the right granularity to be "solved". This again echoes the market process; too many variables, and too many different preferences leads would-be problem solvers to rely on the pretense of knowledge.
Languages combine a host of disparate features, and sometimes simply combining previously known features in novel ways is sufficient iterative improvement. Clojure is one good example of this, where immutability, functional programming, and STM were all known and used to varying degrees, but combining them in a clever way allowed for something greater to emerge, particularly how the first two allowed for a new form of the third. From this in turn emerged a (new?) well thought out model of time state, and identity.
In the end, perhaps languages are now really engineering problems, not the science/math problems they used to be. Perhaps the academics should embrace this. To continue the economic parallel, if one wishes to examine the qualitative aspects of PLs, perhaps research should approach from the historical perspective, asking why certain languages emerged to solve certain problems.