over decades there has an enormous amount of discussion about the rise and fall of various programming languages but somehow it remains for ever a shallow analy

over decades there has an enormous amount of discussion about the rise and fall of various programming languages but somehow it remains for ever a shallow analysis, more noise than signal. what do we really know about this phenomenon at the end of the day? one problem is the difficulty of measuring what we are talking about: when is a language "dead", when the last program in production stops running? when the last programmer that can program in it retires? when the last school that is teaching it removes it from the curriculum? or maybe when the last computer scientist that can enhance it decides there are richer academic and/or commercial pickings elsewhere? another problem is the objective measurement of any of the above definitions. beside the design and availability of surveys, objectivity was always a problem given that software is big business and hence various actors have incentives to skew the perceived state to support their aims. but it has become particularly problematic in the current interconnected universe of developers that leads to very pronounced virality and network effects which may or may not persist. finally the scope of programming is ever expanding as more and more aspects of society and economic life get digital but the process is still in full swing, materially incomplete and dominated by idiosyncratic oligopoly circumstances. think for example about the billion mobile devices around the planet that are currently practically impossible to program by their users. one could imagine e.g., of the "dead" languages (Basic?) suddenly return from the grave as an ubiquitous mobile device programming paradigm. A programming language that was useful in the past in a certain context may become useful again if a similar context reappears. what would be interesting is a sort of "principal component analysis" of the programming language space: identifying the principal functionalities / capabilities provided by different languages. a dead language could then be defined as one that is completely dominated (effectively a subset) of another one along all relevant principal component directions