Clueless academics publishing junk papers at obscure conferences are high priests of the church of programming languages. Dijkstra, (temporary) reached the status close to (false) prophets :-).
On a deep conceptual level building of a new language is a human way of solving complex problems.
Recently I have started working a lot more with various F5 hardware including BIG-IP LTM’s or BIG-IP Local Traffic Managers.
F5 provides a VMware image so you can have a test environment running as a virtual machine.
That means that complier construction in probably the most underappreciated paradigm of programming of large systems.
Much more so then greatly oversold object-oriented programming. For users, programming languages distinctly have religious aspects, so decisions about what language to use are often far from being rational and are mainly cultural.
Or at least avoid it on a particular phase of language development (C is not simpler language then PL/1, but was widely adopted because of the progress of hardware, availability of compilers ).
The second interesting category is number of applications written in parcilar language that became part of Linux or, at least, are including in standard RHEL/FEDORA/CENTOS or Debian/Ubuntu repository.But idea is especially applicable to compilers and interpreters.What Peter Naur failed to understand was that design of programming languages has religious overtones and sometimes represent an activity, which is pretty close to the process of creating a new, obscure cult ;-).A mediocre language with good programming environment can give a run for the money to similar superior in design languages that are just naked. Critical application is also very important and this is a story of success of PHP which is nothing but a bastardatized derivative of Perl (with all the most interesting Perl features surgically removed ;-) adapted to creation of dynamic web sites using so called LAMP stack.Progress in programming languages has been very uneven and contain several setbacks.