Hasty Briefsbeta

All tags

programming languages

29 stories total

Bilingual

Reversing Abstractions: An Existential Crisis
3 days ago
- The author critiques the dominant 'forward/existential' approach in computer science, which hides implementation details, and advocates for 'backward/universal' approaches that make details accessible.
- Examples of opting out of hiding include ABIs (public conventions) and DWARF debuggers (meta-level descriptions), with debuggers representing a universal approach that can recover abstract views from any concretion.
- The neglect of backward/universal methods contributes to software brittleness, contrasting with biological adaptability, and programming language research is particularly averse to this direction.
- Link-time optimization (LTO) currently delays lowering rather than lifting; the author proposes using debugging information to lift object code back to intermediate representation for better flexibility.
more...
Concurrency, interactivity, mutability, choose two
6 days ago
- Concurrency, interactivity, and mutability are three desirable features in programming, but it is impossible to have all three simultaneously without trade-offs.
- Dropping interactivity (as in C, Rust, Go) ensures safety and performance but removes the ability to inspect or modify programs at runtime.
- Dropping concurrency via a Global Interpreter Lock (GIL) in languages like Python and Ruby allows safe interactivity and mutability but introduces performance overhead from locking all data accesses.
- Dropping mutability (as in Erlang) uses message passing with copied values for safe concurrency and interactivity, but copying data is slow and incurs performance costs.
more...
The Shape of Data (2022)
4 days ago
- Consistent data notation across code, logs, debug, and UIs reduces mental load and improves productivity.
- Clojure's edn syntax provides a roundtrip property (read-print) that makes textual representation a direct mental model for data.
- Separating read from eval (like JSON.parse vs eval) ensures safety and enables language-independent data exchange.
- Using tags (e.g., #inst) allows extensible types without sacrificing self-describing notation.
more...
The Hundred-Year Language
5 days ago
- Programming languages evolve like species, with evolutionary dead-ends such as Cobol and Java.
- The core of a language should have few, clean axioms that determine its long-term survival.
- Faster hardware will allow more wasteful but simpler designs, separating semantics from implementation.
- Bottom-up programming, with multiple layers, yields flexible and reusable software.
more...
Nontrailing separators do not spark joy
5 days ago
- JSON's lack of trailing commas complicates adding or removing key-value pairs, requiring different transformations than if trailing commas were allowed.
- Haskell's record syntax uses leading commas, making it easier to modify the last row but harder to change the first.
- TLA+ disallows trailing commas, causing frequent issues when adding variables, while its PlusCal extension does not have this problem.
- Prolog uses a period to terminate rules, not a trailing separator, which can lead to parsing ambiguity if commas were used instead.
more...
Which programming languages are most token-efficient?
5 days ago
- LLMs face context length constraints, making token-efficient programming languages beneficial for AI agents.
- Dynamic languages (e.g., Clojure) are more token-efficient than static ones (e.g., C), with a 2.6x gap.
- Functional languages like Haskell and F# are nearly as token-efficient as dynamic ones due to advanced type inference.
- Array languages like J (ASCII-based) are highly token-efficient, while APL's exotic symbols are poorly tokenized.
more...
Five Questions about Language Design
5 days ago
- Programming languages are designed for people, not computers; they must accommodate human cognitive limitations.
- The best languages are often created by their authors for their own use, avoiding a condescending attitude toward programmers.
- Languages should grant programmers maximum control, as seen in Lisp's macros and openness, treating programmers as equal partners.
- Brevity in code and manuals is highly valued by skilled hackers, and should be a design goal.
more...
The Roots of Lisp
5 days ago
- John McCarthy's 1960 paper demonstrated how a programming language can be built from a few simple operators and function notation.
- Lisp, short for 'List Processing,' uses lists as a core data structure for both code and data.
- McCarthy's discovery is significant as a historical milestone and a model for modern programming trends.
- The C and Lisp models are described as two clean, consistent programming paradigms, with newer languages moving toward the Lisp model.
more...
What Made Lisp Different
5 days ago
- Conditionals (if-then-else) were invented by McCarthy for Lisp and later spread to most languages.
- Functions are first-class objects in Lisp, treated as a data type like integers or strings.
- Recursion was first supported as a programming language feature in Lisp.
- Variables in Lisp are effectively pointers; values have types, not variables.
more...
What Languages Fix
5 days ago
- Kevin Kelleher proposed describing programming languages by the problem they fix.
- Many programming languages can be succinctly characterized in terms of the specific issues they address.
- Examples range from early languages like Algol and Fortran (fixing low-level assembly) to modern ones like Python and Ruby (fixing Perl's kludge).
- The approach reveals how each language evolves to solve perceived shortcomings of its predecessors.
Succinctness is Power
5 days ago
- Succinctness equals power in programming languages; the main purpose of high-level languages is to make source code smaller.
- A better measure of code size is the number of elements (distinct nodes in syntax tree), not lines of code or characters.
- Brooks' observation: programmers produce similar lines of code per day regardless of language, so more succinct languages enable faster development.
- More powerful languages tend to have fewer bugs, which is critical in applications like network switches.
more...
Revenge of the Nerds
5 days ago
- Programming languages are not equivalent; some are more powerful than others, and Lisp is a prime example of a powerful language.
- Managers (pointy-haired bosses) often choose mainstream languages like Java or C++ due to perceived safety and ease of hiring, ignoring the actual power of languages.
- Lisp, created in 1958, introduced many innovative ideas (e.g., conditionals, recursion, garbage collection) that are only now being adopted by mainstream languages.
- Using a more powerful language like Lisp can lead to significantly shorter code (7-20x shorter than C) and faster development, providing a competitive advantage.
more...
If Lisp is So Great
5 days ago
- Popularity of programming languages does not correlate with quality, similar to how Jane Austen's works are less popular than John Grisham's despite higher quality.
- Lisp appears difficult due to its unusual syntax, intimidating many potential users.
- Past excuses for avoiding Lisp, like being too slow, are no longer valid; now the excuse is circular—other languages are more popular.
- Popularity in programming languages is self-perpetuating due to more libraries, easier integration, and manager preferences.
more...
Weird Languages
5 days ago
- The claim that all programming languages are equivalent reflects limited programming experience, referring only to the 99.5% of tasks involving library-function gluing.
- The remaining 0.5% of programming is uniquely interesting and can be explored through 'weird' languages, whose anomalies hint at non-glue programming techniques.
- Lisp macros exemplify such techniques, enabling language creation for specific problem domains and broader program manipulation.
- Learning weird languages, especially those with smart user bases, expands one's concept of programming by focusing on unique features that enable new ways of thinking and expressing solutions.
A Brief, Incomplete, and Mostly Wrong History of Programming Languages
6 days ago
- 1801: Jacquard's loom uses punch cards to weave text, lacking modern features.
- 1842: Ada Lovelace writes the first program without a computer.
- 1936: Alan Turing and Alonzo Church independently invent all future programming languages.
- 1940s: Early computers programmed with wires and switches to avoid syntax debates.
more...
Thoughts on Integers (2023)
11 days ago
- Many popular languages (C, C#, Go, Swift) default to a signed `int` type, biasing programmers away from considering unsigned types.
- Rust treats all integer types equally with explicit sizes (e.g., `i32`, `u64`) and no default `int`, forcing deliberate type selection.
- Rust uses `usize` for sizes and indices, preventing negative array indexes and enabling large arrays, but defaults integer literals to `i32`.
- Signed integer overflow is undefined behavior in C, allowing compiler optimizations like eliding overflow checks, but unsigned overflow wraps.
more...
Online Historical Encyclopaedia of Programming Languages
9 days ago
- The online roster lists 8,945 programming languages from the 18th century to the present.
- It includes 7,800 influence links and over 11,000 citations.
- The genealogical tree of HOPL is maintained by librarians Diarmuid J Pigott and Bruce M Axtens, with quartermaster Steve Poulson.
- The content is copyrighted by Diarmuid Pigott from 1995-2020, with a prohibition on copying or reproduction.
more...
Programming language file extensions that match ISO 3166-1 alpha-2 country code
9 days ago
- ISO 3166-1 alpha-2 provides two-letter codes for countries (e.g., ME for Montenegro) and non-country entities (e.g., EU for European Union).
- These codes appear in top-level domains and emoji flags (e.g., :flag_fm: for Micronesia).
- Many programming languages share their file extensions with country codes (e.g., .rs for Rust and Serbia).
- The article lists examples like AM (Armenia/Automake), ES (Spain/ECMAScript), PY (Paraguay/Python), and RS (Serbia/Rust).
more...
Why care about programming languages
10 days ago
- Programming languages are valuable for the ideas they embody, not just syntax or tooling.
- Concepts like Rust's borrow checker, Haskell's typeclasses, and SQL's relational theory expand your mental toolkit for problem-solving.
- Understanding these ideas helps evaluate AI-generated code, choose appropriate abstractions, and design maintainable systems.
- In the age of AI, deep knowledge of programming concepts distinguishes true engineers from those who only prompt.
A Road to Lisp: Which Lisp
16 days ago
- Lisp is a family of programming languages with many dialects, unlike single-language evolutions like Python or Java.
- Beginners should focus on learning Lisp's programming concepts rather than worrying about which dialect to choose, as switching later is relatively easy.
- Common Lisp (CL) is mature, standardized since 1994, with multiple implementations like SBCL, offering high performance, a powerful REPL, and comprehensive features including CLOS for OOP.
- Clojure is a modern Lisp on the JVM, emphasizing functional programming with immutable data structures, and is widely used in production by companies like Nubank and Netflix.
more...

firstprev0next

programming languages

Reversing Abstractions: An Existential Crisis

Concurrency, interactivity, mutability, choose two

The Shape of Data (2022)

The Hundred-Year Language

Nontrailing separators do not spark joy

Which programming languages are most token-efficient?

Five Questions about Language Design

The Roots of Lisp

What Made Lisp Different

What Languages Fix

Succinctness is Power

Revenge of the Nerds

If Lisp is So Great

Weird Languages

A Brief, Incomplete, and Mostly Wrong History of Programming Languages

Thoughts on Integers (2023)

Online Historical Encyclopaedia of Programming Languages

Programming language file extensions that match ISO 3166-1 alpha-2 country code

Why care about programming languages

A Road to Lisp: Which Lisp