Diversity of programming languages

• Each year at least one programming language is developed
  • from 1947!
  • hundrets of languages
  • but just few of them are used extensively

Paradigmas

• Imperative
  • Procedural
  • Object-oriented
• Declarative
  • Functional
  • Logic
  • Dataflow
• Metaprogramming
• Array-oriented
• etc.

Domain-specific languages

• SQL
  • one of most popular languages
• array languages
  • APL, J
• pattern matching
  • regexp now used everywhere
• specialized query languages
  • GraphQL
  • jq
  • xml path
  • etc. etc.
• data flow languages
• numeric analysis
  • Matlab
  • Julia
• computational science
  • Julia
• statistic
  • R

Abstraction level

• Levels from the real HW
  • 5 applications
  • 4½ (scripting languages, VM-based languages)
  • 4 C-like languages
  • 3 assembler
  • 2 machine code
  • 1 syscalls
  • 0 HW

Technological differences

• Interpreters
• Compilers
• Transpilers
• “Blended” approach

Historical view (simplified)

• FORTRAN
  • ad-hoc compiler
  • semi-procedural language
  • no truly recursive
• Algol
  • compiler based on quite good theory
  • procedural, structured
  • recursive
• Pascal, Modula
  • loosely based on Algol
  • simpler and faster one-pass compilers!

Language popularity is based on

• language design
• syntax, semantic (usually in this order!)
• difficulty to master it
• fashion trends
• what’s used in curricula (secondary schools, uni)
• positive feedback (in both directions)
• ecosystem
• quality of compiler (speed, error messages)

Fashion trends

• functional
• spaghetti code
• structured programming
• message passing
• actor-based
• hybrid OO
• functional (again)

Ecosystem

• some ecosystems are based on just one/two language(s)
• typical example
  • web browsers + JS (or TS)
• transpilers have to be used for other languages

Difficulty to master it

• “number of developers doubles every five years”
  • i.e. in typical team > 50% of developers are juniors
• and they need to fight with tooling, CI/CD, databases too
• -> the easier the language is to grok, the better

Survival of programming language

• appeals to a wide audience
• gets the job done
• fills a niche
• powerfull user base
• a christmatic leader

https://ccrma.stanford.edu/courses/250a-fall-2005/docs/ComputerLanguagesChart.png

Conclusion

• there’s no silver bullet
  • and very probably never be one

It is impossible to have one language to rule them all

• vast syntax
• full of historical baggage after some time
• complicated (and buggy!) compiler
• huge manuals
• less people to master it
• problematic in large companies
• mix of low-level and high-level constructs

Some examples of (too) complicated languages

• PL/I
  • complicated
  • incomplete
  • buggy compilers
• Algol 68
  • 3 forms of program representation!
• C++
  • incomplete compilers for a long time
  • incomprehensible error messages

Counterstrike

• Go
  • designed specifically with junior developers in mind
• Python
  • w/o type specifiers and other modern specialities

Case study - top 5 languages used in Red Hat

• C
• Python
• Java
• Go
• JavaScript

Related presentations

• Top 5 languages used in Red Hat
• Go programming language
• Rust programming language
• Python programming language