Research: problems, properties and open questions
Development status: Research — This note records goals, assumptions and open questions; nothing here is an assurance or a released feature. An experimental implementation exists for internal protocol development, but in it confidentiality and proving features are simulated or switched off, so it does not demonstrate the capabilities at the heart of the aim: confidential processing and results that others can verify. No performance, resilience or audit results are published, no public references are approved yet, and nothing is publicly available for live use.
Before you read: About the Cubane developer documentation.
Where are the problem statements, intended properties, assumptions and open questions? In this section. This page explains what the section contains and leaves out, how research notes are structured and revised, how to read goals as distinct from evidence, and where evaluation results and approved references would appear — neither exists yet.
What this section is for
The Research section brings together the research framing of Cubane: the problem the platform is intended to address, the properties the design is intended to have, the assumptions those properties would rest on, the trade-offs between them and the questions that remain open.
It is written for researchers, and is also useful to developers evaluating fit and to assurance readers, such as auditors and regulators, who need to separate goals from evidence. It assumes working knowledge of distributed systems, cryptography and confidential computation; it does not teach these fields, and it links terms to the Glossary rather than redefining them.
The section states intent and the limits of that intent. How the intended properties would be achieved — the design detail — is not described in this documentation, and no comparisons with other platforms are made. Every page in the section carries the Research label and its phrase: “This note records goals, assumptions and open questions; nothing here is an assurance or a released feature.”
How the section is organised
- Problem statement and intended properties — the problem, the intended aim and each intended property, stated as a goal.
- Assumptions and open questions — a consolidated list of stated assumptions, trade-offs and open questions across topics, each linked to the note or concept page that discusses it.
- Design notes, one per research topic:
- Design note: confidential computation with verifiable results — processing data that stays protected while producing results that others could check.
- Design note: operating under partial failure and evolving participation — degrading gracefully when part of the network is unavailable, and accommodating changes in participation over time.
- Design note: independently checkable randomness — randomness that parties other than its producer could check.
- Design note: selective disclosure and audit records — sharing selected results with entitled parties, and keeping records of actions whose later alteration would be detectable.
- Design note: accountability for misbehaviour — answering misbehaviour on the basis of evidence rather than suspicion.
- Evaluation methodology and published results — no results are published; the page explains what an eventual evidence page would state.
- Public references — no approved references are listed yet; the page explains how references would be added.
A suggested path for researchers:
- This overview.
- Problem statement and intended properties.
- What Cubane is designed to do.
- The security and trust pages, starting with Security model, trust and limitations.
- What remains observable.
- Confidential computation.
- The design notes listed above.
- Assumptions and open questions.
- Evaluation methodology and published results.
- Public references.
- The Glossary.
How a research note is structured
Each design note uses the same fixed sections in the same order, so that notes can be compared with one another and gaps are easy to see. The framing and evidence pages of the section carry the same status header and closing box, but are organised around their own questions.
The sections of a design note are:
- Status header — the Research label and its phrase, the note’s as-of date and its next-review date.
- Motivation — the problem the note addresses and why it matters to developers, organisations or researchers.
- Intended properties — each stated as a goal, never as an assurance, with its own status label where it differs from the rest of the note.
- Assumptions — what the intended properties would rely on, stated qualitatively.
- Limitations — what is not achieved or not shown, including what the experimental implementation does and does not demonstrate.
- Open questions — what remains unsettled or undescribed, phrased as questions.
- What this note does not claim — a closing box listing claims the note deliberately does not make, followed by related pages.
A section with nothing to report says “None stated” rather than being left out, so that an absence is visible instead of hidden.
Reading research material: goals versus evidence
- A property is a goal. An intended property belongs to a design that is planned or being explored. It is not an assurance, and it is not evidence that the property holds.
- The prototype is narrow evidence. The experimental implementation is evidence only for its own limited scope. In it, confidentiality and proving features are simulated or switched off, so it provides no confidentiality protection; see Known limitations of the experimental prototype.
- No measured results. No performance, resilience or audit results are published, so nothing in this section should be cited as a measured result.
- Absences are stated. Where something is not described, the notes say so plainly. Do not infer a preferred answer from the way a question is phrased.
- Mixed maturity is marked. The labels used on other pages — Experimental implementation exists, Planned and Being explored — describe the maturity of individual capabilities. A research note can discuss capabilities of different maturity and marks each one. Documentation conventions: status labels, dates and reviews explains the labels.
How notes are reviewed, revised and superseded
- Each note carries an as-of date and a next-review date, and notes are reviewed periodically.
- A substantive revision updates the as-of date and is recorded in the Documentation change log.
- A superseded note stays published at the same address with a “not current” banner linking to its replacement, so that existing citations still resolve.
- New notes may be added over time. If a note is withdrawn, its address leads to this overview with a neutral note.
- Status changes for any intended property are dated and recorded in the change log.
Citing this material
- No approved public references, such as papers or specifications, are listed yet; see Public references.
- When citing a research note, cite its title, its address and its as-of date, and present its content as goals and open questions as of that date.
- Do not cite the behaviour of experimental software, or any material that is not listed on Public references, as Cubane results.
Assumptions, trust and limitations
- Research notes are design-stage material. Their content may change, be deferred or be withdrawn, and none of it is a commitment.
- Topics without a note are not described in this documentation; that absence implies nothing about their status.
- Nothing in this section is a security, confidentiality, correctness or availability assurance.
Not yet available, and how this page will grow
- Approved public references would be listed on Public references once approved; none are listed yet.
- Published, versioned evaluation results would be listed on Evaluation methodology and published results; none are published yet.
- New notes would be added to this overview with their as-of dates, and superseded notes would keep their addresses with a banner.
Related pages
- Problem statement and intended properties — the starting point of the research framing.
- Assumptions and open questions — the consolidated list of assumptions, trade-offs and open questions.
- Evaluation methodology and published results — how evaluation would be reported; no results are published.
- Public references — approved references and how they would be added; none are listed yet.
Glossary
References
No approved public references yet.
