Vision
A future in which people can transform data and ideas into trustworthy knowledge through computing that is accessible, understandable, and accountable.
The vision describes the future Aner seeks to support.
Mission
To build an open source programming language and ecosystem that simplify data science, scientific computing, and AI without sacrificing performance or understanding. Aner aims to help people create, verify, reproduce, and scale computational work, with transparency, traceability, and responsible human and AI collaboration as foundational commitments.
The mission explains what Aner will build to help achieve its vision.
The 10 Founding Principles of Aner
1. AI native, human led
AI should expand human capability while preserving human judgment, agency, and accountability.
Aner should support both the creation of AI systems and collaboration with AI during programming, analysis, and research. However, AI should remain an option, not a requirement for using the language.
Whether code is written by a person, an AI system, or both, people should be able to inspect, test, modify, and control its behavior. AI generated code should meet the same standards of correctness and responsibility as human written code.
Design commitment: Make AI a first class capability without making it an unquestionable authority.
2. Simplicity with clarity
Make computational ideas easier to express, not merely shorter to write.
Aner should favor readable syntax, consistent concepts, understandable errors, and a small number of well designed abstractions. It should help students, researchers, and practitioners move from an idea to a working implementation without unnecessary complexity.
Simplicity must not conceal important assumptions, numerical choices, resource costs, or side effects. Convenient defaults should remain discoverable and adjustable.
Design commitment: Reduce unnecessary complexity while preserving the information people need to understand their work.
3. Transparency by design
Important computational behavior should be visible, inspectable, and explainable.
Users should be able to understand how Aner interprets their instructions, executes operations, selects computational resources, and interacts with external systems. Automatic optimizations and execution decisions should be inspectable rather than mysterious.
Transparency also applies to the project itself: documentation should distinguish established capabilities from experimental features and future proposals. Limitations and performance claims should be communicated honestly.
Design commitment: Explain what the system does, expose consequential decisions, and state limitations clearly.
4. Traceability and meaningful provenance
A result should have a documented path back to its origins.
Aner should make it possible to connect outputs with the data, transformations, code versions, models, and execution settings that produced them. When people and AI collaborate, their contributions should be documented through declared or captured provenance: not assumed from the appearance of the code.
Traceability should distinguish verified information, user provided declarations, and missing history. It should preserve useful evidence without unnecessarily exposing sensitive information.
Design commitment: Help users answer, “Where did this result come from, what changed along the way, and what evidence supports that history?”
5. Reproducibility and scientific integrity
Computational work should be open to examination, testing, and meaningful repetition.
Aner should help users preserve relevant assumptions, data context, units, dependencies, configurations, and sources of randomness. Results should be accompanied by enough context to support verification and interpretation.
Where exact repetition is not supported, the system should make relevant sources of variation visible rather than promise identical outcomes. Uncertainty and numerical limitations should be communicated, not hidden behind a precise looking answer.
Design commitment: Make verification and reproducibility part of the workflow, rather than an afterthought.
6. Ethics and responsibility for consequences
Technical capability does not remove responsibility for how it is used.
Aner should encourage users to consider the effects of their systems on individuals, communities, and the environment. Its ecosystem should support practices such as evaluating bias, documenting intended uses and limitations, examining potential harms, and identifying who is accountable for consequential decisions.
Ethics should not be reduced to a label or a single automated score. Aner should support responsible judgment without claiming that a programming language can guarantee ethical outcomes.
Design commitment: Help people evaluate and document consequences, not simply optimize technical objectives.
7. Privacy and security by default
Trustworthy computation must protect the people, data, and systems involved.
Aner should favor safe defaults, explicit permissions, controlled access, and careful handling of sensitive information. Connections to AI providers, external services, and distributed resources should make data movement and permissions understandable.
Privacy must also shape transparency and traceability. Audit records should not expose credentials, personal information, proprietary data, or private prompts merely to make a workflow easier to inspect.
Design commitment: Collect and expose only what is necessary, and keep users in control of sensitive information and privileged actions.
8. Performance with resource responsibility
Powerful computation should use resources deliberately, efficiently, and transparently.
Aner should pursue efficient execution across CPUs, GPUs, and distributed systems while helping users understand tradeoffs involving speed, memory, numerical accuracy, and resource consumption.
Its design should also address workloads that exceed available memory, rather than assuming that every dataset fits into RAM. Performance improvements should be demonstrated with reproducible evidence, not unsupported claims.
Design commitment: Make scalable computing accessible without hiding its costs or compromising correctness.
9. Openness, interoperability, and user freedom
The ecosystem should expand users' choices rather than restrict them.
Aner should remain open to inspection, contribution, extension, and integration with other languages, libraries, data formats, and computational systems.
Users should retain control over their code, data, models, and execution environments. Core language use should not depend on a mandatory AI provider or cloud service, and interoperability should help people adopt Aner without unnecessarily abandoning existing work.
Design commitment: Build an open ecosystem that people can use, extend, and integrate without avoidable lock in.
10. Community stewardship and fair recognition
Aner should grow through contributions that are welcomed, evaluated fairly, and credited honestly.
The project should encourage participation from researchers, developers, educators, students, and practitioners with different backgrounds and levels of experience.
Design decisions should be documented, governance should be understandable, and disagreements should be addressed through respectful discussion and evidence. Recognition should include code, research, documentation, testing, teaching, and maintenance: preserving original authorship while giving subsequent contributors accurate and visible credit.
Design commitment: Build a community in which responsibility is clear, participation is meaningful, and credit reflects actual contributions.
Distinct Responsibilities That Reinforce One Another
Transparency, traceability, reproducibility, and ethics are connected, but they answer different questions. Aner should address each explicitly rather than treating one as a substitute for the others.
| Foundation | Question Aner should help users answer |
|---|---|
| Transparency | What is the system doing, and what assumptions or decisions affect its behavior? |
| Traceability | Where did this result come from, and what happened along the way? |
| Reproducibility | Can someone meaningfully repeat or independently check this work? |
| Ethics and accountability | Should this system be used this way, who may be affected, and who is responsible? |
A traceable result is not automatically correct, and a reproducible result is not automatically ethical.
Aner’s purpose
Aner is being developed as an open source programming language for data science, scientific computing, and AI, guided by simplicity, performance, and responsibility. Its purpose is to help people express ideas clearly, scale computation efficiently, and preserve the context and evidence needed to understand, verify, and responsibly use their results.
A public software release is planned. The software license and contribution process will be finalized before launch. See the release status for current availability.
Guiding words
Clear intent. Traceable computation. Responsible AI.
These words express Aner’s direction and values. They are not a certification or guarantee of technical capabilities or ethical outcomes.
Core founding commitment
Aner should make it easier not only to compute an answer, but to understand where it came from, evaluate whether it deserves confidence, and take responsibility for how it is used.