Overview of programming languages
Programming languages span diverse paradigms, formal semantics, and typing disciplines, with modern development dominated by a small set of widely applicable languages such as Python. A prominent contemporary trend is the growing role of generative artificial intelligence and code-writing assistants, which handle lower-level syntax and shift programmer attention away from language-specific particulars. Consequently, public question volumes on developer forums have sharply declined, and the barriers to establishing new programming languages have increased.
Foundational Concepts and Paradigms
Programming languages can be classified into major families, including imperative, object-oriented, functional, and declarative paradigms [1]. Within and across these language families, implementations differ across design dimensions such as dynamically typed versus statically typed systems, call-by-name versus call-by-value evaluation strategies, and compiled versus interpreted execution models [1][2]. The theoretical foundation of programming languages also encompasses formal semantics, which include operational, denotational, and axiomatic semantics [3].
A fundamental distinction in programming paradigms exists between value-level programming and function-level programming, which John Backus identified in his work on programs as mathematical objects [4]. In contrast to value-level definitions where program variables are essential, a function-level program is variable-free and point-free [5]. The canonical language designed for function-level programming is FP, with others including FL and J [6]. To advance language design and programming methodology from an informal trade-craft into a rigorous discipline, foundational frameworks require establishing precise definitions [7]. In academic research, investigators develop mathematical descriptions of programming languages to prevent programmer mistakes [8].
Historical Evolution and Language Design
Historically, programming languages provided a way for humans to describe computer processes, as reflected in early languages such as FORTRAN and BASIC [9]. As computers and descriptive methods became more sophisticated, programming languages began imitating natural languages in their look and feel to reflect how people think [10]. However, programming languages remain far more precise than natural languages, leading to ongoing work translating precision-demanding legal text into programs [11].
New programming languages frequently build on older predecessors, often marketing themselves as "C++ but better" or "Python with X" [12]. Developers often create new languages to address perceived problems in older ones or to test new ideas, with successful languages providing a new niche while remaining widely applicable [13]. Despite such experimentation, Backus criticized the tendency of successive languages to incorporate predecessor features with superficial cleaning up, arguing that few new languages make programming sufficiently cheaper or more reliable [14]. Moreover, contrary to the myth that writing code is the hardest part of software development, most effort is typically spent on system design and human management [15]. In practice, a small set of languages continues to be used for the vast majority of new applications [16].
Measurement and Rankings of Language Popularity
Efforts to gauge programming language popularity focus on languages in active use, including activity from maintaining legacy systems [17]. For instance, IEEE Spectrum evaluates 64 programming languages by weighting multiple data sources to construct composite popularity indexes [18]. Its methodology tracks three separate dimensions of popularity: working use among software engineers and typical IEEE members in the Spectrum ranking, employer demand in the Jobs ranking, and zeitgeist relevance in the Trending ranking [19]. Employer demand is captured through job listings on portals such as the IEEE Job Site [20]. Because language names like C++ and Scheme frequently collide with common terms in research literature, and because APIs are subject to changes and terminations, metrics data collection has shifted toward manual gathering [21].
In the 2025 rankings, Python secured the top position in the default Spectrum ranking, whereas JavaScript dropped from third place to sixth place [22]. Python also achieved first place in the employer-focused Jobs ranking, rising from second place the previous year, while SQL expertise maintained strong value on developer resumes [23].
Artificial Intelligence and Changing Developer Trends
The emergence of code-writing artificial intelligence is shifting how software is written, as automated assistants increasingly handle programming layers starting with syntax and extending through flow control and functions [24]. Given sufficient training data, language models can generate code for prompts in any requested language, which in practical terms encourages relying on today's most popular general-purpose programming languages [25]. Conversely, programmers report that artificial intelligence models deliver noticeably poorer results when generating code in less-used languages [26].
As a result, it is becoming significantly more difficult for new programming languages to emerge and achieve adoption [27]. At the same time, developer reliance on public forums is dropping sharply as coders consult private language models and coding assistants; across evaluated languages, weekly questions on Stack Exchange in 2025 dropped to 22 percent of their 2024 level [28].
Key facts
- Programming languages span imperative, object-oriented, functional, and declarative paradigms, along with operational, denotational, and axiomatic formal semantics [1][3].
- Function-level programming, represented canonically by FP, FL, and J, eliminates program variables to operate in a variable-free mathematical space [4][5][6].
- Early languages like FORTRAN and BASIC described computer processes, whereas later designs evolved to approximate natural language syntax and cognition [9][10].
- New languages frequently market themselves by building on established tools like C++ or Python, yet the majority of new applications remain concentrated in a small set of languages [12][16].
- Writing code is rarely the most demanding part of software development; design and human management consume the primary effort [15].
- Python led both the IEEE Spectrum general ranking and the employer-focused Jobs ranking in 2025, while JavaScript dropped from third to sixth place [22][23].
- Code-writing artificial intelligence tools handle syntax, flow control, and functions, but perform noticeably worse on less-used programming languages [24][26].
- Across tracked languages, weekly question volume on Stack Exchange in 2025 fell to 22 percent of the 2024 level as developers turned to conversational models [28].
- The dependence of artificial intelligence models on large training datasets makes it harder for newly created programming languages to gain traction [25][27].
Sources
-
2027-28 - COMP2322 - Programming Language Concepts | University of Southampton www.southampton.ac.uk
- [1]
The differences between families of languages (imperative, OO, functional, declarative) and within families (dynamically typed vs statically typed, call by name vs call by value, etc)
- [2]
Compiled vs. interpreted languages
- [3]
Programming language semantics: operational, denotational and axiomatic semantics
- [1]
-
Function-level programming - Wikipedia en.wikipedia.org
- [4]
In computer science, function-level programming refers to one of the two contrasting programming paradigms identified by John Backus in his work on programs as mathematical objects, the other being value-level programming.
- [5]
A function-level program is variable-free (cf. point-free programming), since program variables, which are essential in value-level definitions, are not needed in function-level programs.
- [6]
The canonical function-level programming language is FP. Others include FL, and J.
- [14]
Each successive language incorporates, with a little cleaning up, all the features of its predecessors plus a few more. [...] Each new language claims new and fashionable features... but the plain fact is that few languages make programming sufficiently cheaper or more reliable
- [4]
-
Practical Foundations for Programming Languages www.cs.cmu.edu
- [7]
If language design and programming methodology are to advance from a trade-craft to a rigorous discipline, it is essential that we first get the definitions right.
- [7]
-
Talking computer languages with Cameron Wong seas.harvard.edu
- [8]
My work is in developing mathematical descriptions of programming languages, and then using those descriptions to prevent programmer mistakes.
- [9]
Historically, programming languages have been a way for humans to describe computer processes, and early languages like FORTRAN or BASIC really reflect this.
- [10]
As computers (and the way we describe them) have gotten more sophisticated, though, we've started seeing programming languages try to imitate natural languages in their look and feel, since that more closely reflects how people think
- [11]
Programming languages are much, much more precise than natural languages, so much so that there is ongoing work to translate legal text (a setting where precision is demanded and necessary) to be described as programs.
- [12]
All the time. It's pretty common to see new languages market themselves as "C++ but better", or "Python with X.”
- [13]
People often make new languages to try to fix perceived problems with older ones, or to experiment with new ideas, and the successful ones are those that both provide a new niche and are also widely applicable.
- [15]
Probably the most pervasive myth about programming in general is that the hardest part of software development is actually writing the code -- in most cases, the real effort is spent on design, human management, and so on.
- [16]
There's definitely a small set of languages that get used for the majority of new applications these days.
- [8]
-
Top Programming Languages Methodology 2025 spectrum.ieee.org
- [17]
By popularity, we mean we are trying to rank languages that are in active use, including activity from maintaining legacy systems.
- [18]
In total, we identify 64 programming languages. We then weight each data source to create an overall index of popularity, excluding some of the lowest scorers.
- [19]
We look at three different aspects of popularity: languages in active use among typical IEEE members and working software engineers (the “Spectrum” ranking), languages that are in demand by employers (the “Jobs” ranking), and languages that are in the zeitgeist (the “Trending” ranking).
- [20]
We measured the demand for different programming languages in job postings on the IEEE Job Site.
- [21]
In the past we relied heavily on APIs to gather data from sources, but now the data is gathered manually to the difficulty of keeping up with API changes and terminations, and because many of the programming languages’ names (C++, Scheme) collided with common terms found in research papers
- [17]
-
AI Is Redefining the Concept of a Programming Language's Popularity spectrum.ieee.org
- [22]
In the “ Spectrum ” default ranking, which is weighted with the interests of IEEE members in mind, we see that once again Python has the top spot, with the biggest change in the top five being JavaScript ’s drop from third place last year to sixth place this year.
- [23]
in the “Jobs” ranking, which looks exclusively at what skills employers are looking for, we see that Python has also taken fir st place, up from second place last year, though SQL expertise remains an incredibly valuable skill to have on your resume.
- [24]
First details of syntax, then flow control and functions, and so on up the levels of how a program is put together—more and more is being left to the AI.
- [25]
As long as there’s enough training data, they’ll generate code for a given prompt in any language you want. In practical terms, this means using one—any one—of today’s most popular general purpose programming languages.
- [26]
Consequently, programmers have noted that AIs give noticeably poorer results when trying to code in less-used languages.
- [27]
One obvious long-term consequence to this is that it will become harder for new languages to emerge.
- [28]
For example, across the total set of languages evaluated in the TPL, the number of questions we saw posted per week on Stack Exchange in 2025 was just 22 percent of what it was in 2024.
- [22]