A statement from Input | Output Research (IOR) in respect to the Cardano Roadmap Tier List recently published by the Cardano Foundation

[fergiemiller]

The Cardano Foundation (CF) has published a Cardano Roadmap Tier List following an initial roundtable between Giorgio Zinetti (CTO, CF), Sebastien Guillemot (dcSpark, Paima Studios) and Matthias Benkort (Technical Director, CF) on the 28th March. A further roundtable was recently organised by CF on the 22nd April to continue the discussion.

It is great to see community initiatives like these and IOR welcomes discussion and debate from all actors in the Cardano ecosystem, particularly those which we wish to engage with more collaboratively and who are as notable as CF. IOR acknowledges these types of initiatives are much needed and we are working hard ourselves to improve how we engage with the Cardano community.

At the start of both roundtable discussions, it was stated that it is a fun TLDR exercise with people expressing their own opinions. However, we are seeing the tier list slowly becoming perceived as an official statement from CF and we would like to raise some observations we have made into the process and outcome (aka the Tier List) in light of this.

IOR Observations

The information used in the analysis for creating the Tier List was taken from several sources. One of the main sources was the Intersect Draft Bucket Budget Google sheet which was intended as a financial document, rather than a technical or scientific one.

The items grouped together include Research items proposed from IOR via the Intersect Product Committee, Engineering items proposed by the Intersect Technical Steering Committee, several items from the Intersect Product Committee’s roadmap, as well as from a few other sources.

The background of the evaluators in recent discussions has been weighted towards those with an engineering background, with several leaders of software labs within the ecosystem invited to take part. All feedback from technical/engineer community members is valuable. However, as far as we are aware, no one with a formal research background was involved in the tier list evaluation process. So we’d like to clarify a few points we think are important when assessing the relative merits of the different line items.

Software Readiness Timelines

It is important to recognise that research is a different discipline which operates on different time horizons to engineering. Maturing capabilities in emerging technologies and complex systems towards market readiness require considerable efforts that spans over multiple years.

The Software (also called Technology) Readiness Level was a guide consisting of 9 levels which was first developed by NASA, and has since been adopted by organisations such as the European Union, which outlines a development pathway with guidance on best practice.

Late stage SRLs 6-9 are product-orientated and near market, ideally with 12-18 month timelines, mid-stage SRLs 3-5 focus on innovative foundational engineering are 2-3 years out from mainnet deployment, whilst research functions at early SRLs 1-2 with a 3-5 year time horizon.

Engineering-Weighted Evaluation

When looking at the rankings from CF tier list, alongside those from other community driven initiatives (which we also applaud btw) such as the Cardano Roadmap Budget Workshop in Tokyo held in March, we notice a systematic bias in favour of product and engineering items over research.

It is perhaps not surprising that if you evaluate research items through an engineering lens, which considers near term impact, that research will rank poorly.

Yes, tactical short term research initiatives can support product and engineering to solve individual use cases. However, foundational research, by its nature, is not designed to solve challenges over a 12-18 month horizon. Foundational research solves tomorrow’s problems, not today’s, which we would argue requires a different scorecard to that of engineering.

Research Evaluation Timeline

So, how should you evaluate research over these longer time horizons? 5 years is a long time to wait, and could potentially be very costly, particularly if nothing comes of the research.

After all, early stage endeavours carry a high degree of risk. There are lots of unknowns, lots of things that can change (the market certainly will), and in many instances you are literally starting with a short problem statement. How do you predict the future?

You make bets. Strategic ones. Leading corporations do. Countries do. They all invest into scientific programmes. Like Cardano has done.

A scientific vision conducts foundational research in areas that, to the best of your knowledge, you believe the market will be in 3-5 years, and that are typically broadly applicable.

You do this, knowing that a number of these bets won’t come off, but the ones that do will create a source of value that is highly defensible, very hard to copy, provides competitive advantage and exceptional ROI. Ouroboros - one of the most cited and respected consensus protocols in the world - being a case in point.

A Research Evaluation Scorecard

Given this, what criteria should you use to evaluate research? Here are some suggestions from IOR:

Track record - what has the research group done before, what has been their impact?

Inputs - what is the team, what skills do they have, what networks are they in and what resources do they bring?

Proposed work - no-one can predict the future but does the vision, and the items under it, sound sensible and exciting?

Planning - do the project plans sound reasonable? Can the tasks be achieved? Are the deliverables realistic?

Flexibility - things change. Can the plan? Can it stay current and increase its chance for success?

Accountability - are there suitable checks and balances?

Impact - if a number of these research items are realized, what impact will they create?

IOR’s request to the community

We’re now entering a period where DReps will assess the many proposals put forward within the budget process. We’d encourage all DReps to consider the scorecard above as part of that assessment, or indeed any other they may choose.

It’s important that our proposal is assessed through a research lens - that recognises medium- and long term impact, not a product or engineering lens that looks within a much shorter time horizon.

If you are unfamiliar with evaluating research, or have any questions or concerns, please do reach out to us. We’d love to jump into calls and X Spaces with you.

We have a vision for the next 5 years, one that we believe will enable Cardano to lead in blockchain, and create the impact that’s much needed in the world.

We believe this vision is truly exciting. We sincerely hope you do too.

[fergiemiller]

Further clarifications on the IOR proposal

Further to our note above, IOR wishes to provide some additional comments on individual workstreams critiqued by leaders in the community.

The IOR proposal on govtool summarising our proposal at a high level and includes the following supporting links:

• IOR written proposal - an Executive summary supported by 3 sections detailing the Strategic Research Agenda, Cardano Vision and Work Program 2025. This is a comprehensive proposal document providing a 360 degrees of the IOR proposal to a high level of transparency.

• Fundamental Research Streams Google sheet - itemised details of the 20 research streams recommended for prioritisation in 2025

• Technology Validation Streams Google sheet - itemised details of the 6 streams recommended to be brought from SRL2 to SRL5 for implementation by an engineering team from the community

• Cardano Vision presentation - a digestible visual overview of the IOR proposal including details of IOR’s approach and a high level perspective of the five-year vision across 9 thematic or research areas

Whilst we have put considerable effort into the proposal and worked over the last 6 months to shape this with the Intersect Product Committee and other stakeholders, we recognise that not everyone has been able to engage and discuss the proposal in this way.

We have therefore provided some additional clarifications and supporting information below. We are used to peer review in research . We are also open to discussing any of these items publicly so please comment below or invite us to an online meeting or X Space. We’d be happy to join.

Specific stream feedback posted on Govtool by @sebastiengllmt, @cerkoryn and others:

State-Machine Contract Environment (WOS-2)

“I don’t believe it’s worth doing more research on the current EUTXO model”

The EUTXO model remains a core innovation of Cardano’s mainchain and is increasingly vital as we progress toward greater layer 1 scalability, especially with developments like Leios. Specifically, the EUTXO model uniquely allows SPOs to process transactions concurrently in separate input blocks and efficiently reconcile any divergent protocol states once these blocks are serialized on-chain.

For more complex contracts, formal analysis and verification become essential to confidently deploy these contracts. The state-machine approach, initially developed within earlier Hydra research, proved highly effective for handling complex off-chain state. Given its potential, we’ve spun it out into a dedicated research stream, aiming to provide the rigorous formal verification needed to unlock greater scalability and reliability for Cardano’s layer 2 solutions.

Location-Based Services and Smart Contracts (WOS-6)

“I think this is an engineering challenge and not a research challenge”

There are certainly important engineering aspects to address here, but we would argue that fundamental research remains essential. For instance, recent work by Kohls and Diaz (presented at Usenix Security ’22) on the Verloc system demonstrates early efforts towards achieving verifiable geographic localization in distributed systems—but this is only a starting point, and significant open research questions still remain.

Our roadmap for location-based services on Cardano involves incentivizing geographic decentralization with verifiable proofs, enabling smart contracts to leverage reliable geographic information. Location information is becoming increasingly critical for blockchain ecosystems, impacting compliance, access control, and broader user adoption of DApps. Reliable geographic data can significantly enhance protocol participation and verification accuracy, ultimately boosting trust and reliability in decentralized applications.

Proofs of Useful Work (OO-6)

“ I don’t think the value for Cardano is worth the cost”

Exploring PoUW aligns with Cardano’s vision of eventually expanding beyond a purely coin-based Proof-of-Stake model. As demonstrated by our research with the Minotaur protocol, integrating multiple resource types—including computational resources—can significantly enhance ledger security and reflect diverse forms of stakeholder value. By incorporating useful computational tasks like SNARK generation into blockchain operations, Cardano can create new opportunities for SPOs to generate additional income and increase network utility.

Adopting PoUW would notably boost the computational power securing the blockchain, whilst PoUW-based blockchains could attract valuable industrial partners into the Cardano ecosystem, particularly companies seeking cost-effective solutions to recurring optimization problems (e.g., logistics or transportation optimization).

Decentralized Identity and Reputation (GI-1)

“​​I think this is an engineering challenge and not a research challenge.”

While decentralized identity involves significant engineering, there’s a fundamental research layer that’s equally critical. Recent innovations like ZKlogin (introduced at last year’s ACM CCS conference and now integrated as a prominent feature in the Sui blockchain) highlight that embedding user identities into blockchain addresses involves more than just straightforward implementation. Our current research specifically addresses how we can maintain essential properties—like privacy and Sybil resistance—in decentralized identity systems, challenges that inherently require deeper theoretical exploration beyond pure engineering.

From an impact standpoint, formalizing decentralized identity systems and related frameworks such as PKIs will significantly strengthen Cardano’s infrastructure. By combining theoretical results and proof-of-concept implementations, this research ensures that identity integration into Cardano-based services happens securely and in a privacy-preserving manner, substantially improving trust, usability, and resilience across the entire ecosystem.

Tokenomics (TO-1, TO-2)

“Empirically, previous tokenomics research has been the equivalent of burning money in a fire.”

It’s understandable that there’s skepticism around tokenomics, particularly given the current state of the art which is admittedly poor. However, economics and mechanism design are rich, sophisticated fields—when approached with rigour, tokenomics research can provide substantial value and critical insights. Recent work by IOR underscores the need for nuanced, mathematically-grounded approaches to tokenomics—emphasizing rewards, staking mechanisms, and the development of the relevant theoretical underpinnings that distinguish Cardano from other platforms.

To maintain this competitive advantage, we argue that continued exploration and refinement are essential, particularly given that competitors like Polkadot are pursuing similar tokenomics goals—albeit without a clear roadmap, risking issues like “super-inflation spirals.” Ethereum’s rapid progress through its EigenLayer initiative also emphasizes how critical focused, strategic investment in tokenomics has become. Cardano has the potential to not only avoid falling behind, but can actively lead innovation in tokenomics.

Next-Level Governance Protocols (D4-1)

“This is too vague”

​​Currently Cardano’s governance relies heavily on a number of Delegated Representatives (DReps), which inherently restricts scalability and direct stakeholder participation. Our research aims to enhance governance scalability and inclusivity by using layer-2 techniques, making it possible to handle larger-scale participation and more complex decision-making processes efficiently. Additionally, we aim to address critical issues like voter privacy and coercion resistance, enabling secure and trustworthy voting mechanisms at both the DRep and direct stakeholder levels. The goal is to create clearer, more effective, and broadly accessible governance protocols, significantly improving the Cardano ecosystem’s capacity to handle decentralized decision-making at scale.

Hydra (IHT-1, IHT-2, IHT-3, IHT-4)

“I think Hydra is no longer a research item. Either it succeeds in productization, or it dies”

From research’s perspective, “Hydra” has always represented a broader research framework for exploring layer 2 scaling solutions on Cardano. There’s significant value in continuing research to fully map out the design space, investigating variants of Hydra Heads (e.g., allowing dynamic node participation or partial failures) and Hydra Tails (integrating incentives and zero-knowledge proofs for optimal optimistic responsiveness in a client server setting).

This ongoing research isn’t limited to Hydra’s direct implementation—it also directly informs and enhances parallel layer 2 initiatives such as Midgard and Gummiworm. This means Hydra research has wider strategic value, ensuring Cardano remains adaptable, robust, and well-positioned to meet future scaling challenges.

[fergiemiller]

Specific stream feedback from the Cardano Roadmap Tier List published by the Cardano Foundation

Light client infrastructure (Research, IC-3)

“It’s too vague, lacks clear direction and potentially redundant with existing solutions like Mithril”

The Cardano ecosystem lacks a foundational light client infrastructure capable of supporting advanced use cases like trustless zk-bridges, smart contract interactions, and secure DApp integration. While Mithril provides basic cryptographic state proofs, it is just one piece of the puzzle; a wide array of additional requirements are needed such as decentralizing the state storage, working out the incentives around light client support and mapping the entire use case space for all types of applications. This proposal seeks to address those gaps with a future-proof framework optimized for asymmetric data access, device constraints, and robust incentive mechanisms.

Unlike simpler transaction-verification tools, the proposed light client architecture is designed for complex, trust-minimized interactions—critical for cross-chain interoperability and Cardano’s broader strategic goals. A dedicated, research-driven approach is essential, as incremental enhancements to Mithril or tooling-focused proposals are insufficient for developing secure, long-term blockchain infrastructure. We are proposing that this is advancing through technology validation in parallel to ongoing research to speed up implementation.

Hydra Tail (IHT-2)

“Hydra Tail is currently unfeasible due to immature zero-knowledge infrastructure and underwhelming prototype performance, making any near-term implementation premature.”

Hydra Tail enhances Cardano’s Layer 2 scalability by integrating zk-rollups into the Hydra framework. Unlike the peer-to-peer focus of Hydra Head, Hydra Tail enables trustless scaling for a broader client base by batching off-chain transactions and posting succinct zero-knowledge proofs to the main chain. This approach reduces on-chain load while maintaining strong security guarantees and censorship resistance, allowing users to seamlessly move assets and contracts across layers.

This research stream aims to proactively develop the zk-rollup infrastructure needed for Cardano, focusing on formal proofs, efficiency, and long-term ecosystem readiness. While early prototypes have exposed performance challenges, the proposal emphasizes rigorous analysis and phased development to ensure scalability and security. StarStream and ZK-VMs are seen as complementary, not replacements—zk-rollups here primarily aim to prove correct execution, not necessarily conceal any computation, and thus could come with less overhead. By investing in foundational cryptographic work now, Cardano can position itself to lead in the adoption of secure, scalable Layer 2 solutions as the technology matures as outlined at the top of this post.

Fair transaction processing (OO-3)

“While fair transaction processing addresses important inequalities during congestion, the proposal remains exploratory, lacks clear technical implementation plans, and thus isn’t currently actionable or a priority.”

Like Bitcoin and Ethereum, Cardano’s Ouroboros currently lacks protections against front-running and preferential transaction ordering, leaving users vulnerable to practices like Maximal Extractable Value (MEV). While approaches such as proposer-builder separation may mitigate MEV, they often introduce centralization risks at the Layer 1 level. This proposal focuses on building protocol-level solutions to ensure fair transaction processing by default—protecting users, enhancing equity, and maintaining decentralization and performance.

The stream aims to replace fee-based prioritization with more equitable transaction ordering, benefiting real-time applications and reducing market distortion. It addresses complex issues like batch Layer 2 submissions, ensuring fairness across transaction types, and includes formal analysis of multiple implementation paths. With compatibility planning for Ouroboros Leios and peer-reviewed foundational work already in place, this research is not just exploratory—it targets actionable protocol enhancements. It also includes incentive compatibility analysis to strike a balance between fairness and economically driven behavior, with the goal of timely deployment alongside Leios advancements.

[fergiemiller]

Specific stream feedback from the Cardano Roadmap Budget Workshop in Tokyo in March

Multi-resource consensus - Minotaur (OO-5)

“The Minotaur multi-resource consensus approach, while academically interesting and potentially useful for sidechains or L2s, lacks relevance for Cardano mainnet, faces limited practical demand, and therefore isn’t a suitable priority for Cardano treasury funding.”

Minotaur explores multi-resource consensus mechanisms—combining proof of work (PoW), proof of stake (PoS), and proof of restake (PoRS)—to boost blockchain security and resilience. This approach is particularly valuable for launching new blockchains with limited liquidity or uneven stake distribution, leveraging restaked PoS to ensure early stability and enabling future independence.

The initial Minotaur paper focuses on layer-1 protocols, but a key use-case is supporting the launch of partnerchains. This aligns with Cardano’s strategic goal of becoming a secure base layer for sidechains and rollups, expanding its ecosystem and attracting external developers and liquidity. By incorporating restaking (PoRS), Minotaur reflects modern blockchain trends and strengthens Cardano’s position in the growing sidechain market. Treasury support for this work directly promotes DAOs, ecosystem growth, innovation, and long-term value for the Cardano platform.

RSnarks (TV-4)

“RSnarks), while conceptually valuable, are already being pursued by Midnight and StarStream, reducing the urgency and novelty for Cardano’s main chain, and shifting the focus from foundational research toward practical integration.”

This workstream focuses on recursive SNARKs, specifically tailored for Cardano, enabling scalable and privacy-preserving features such as trustless zk-bridges between Cardano and partnerchains. It builds upon Halo2 and KZG commitments to generate compact, efficient proofs. The primary innovation lies in translating Halo2 proofs on Midnight and other partner chains into Halo2-BLS proofs compatible with Cardano’s Plutus environment, a process that demands the development of a recursive prover, a Plutus-compatible verifier, and advanced tooling for cross-curve arithmetic.

Although other projects like Midnight explore recursive SNARKs, this stream uniquely addresses the specific technical constraints and optimizations required by Cardano’s EUTxO model and Plutus infrastructure, ensuring seamless integration rather than the superficial compatibility of other tools. By advancing foundational methodologies—particularly cross-curve translations—this research fills critical gaps not addressed elsewhere and establishes standards for future ecosystem-wide coordination and integration.

Investing early in this foundational research is strategically urgent for Cardano’s competitive positioning, and the scope for this stream is to validate the technology to SRL4, ensuring it is fully validated and ready for implementation. As recursive SNARKs rapidly become fundamental to blockchain scalability and interoperability, this investment ensures Cardano remains at the forefront, unlocking powerful future user-facing innovations.

[Arnaud_Bailly]

Thank you for clarifying your position and the very detailed motivation for each of those research streams.

I have been a participant to one of those “Tier List” roundtables, and possibly was too vocal with my own biases towards focusing on concrete improvements in the developers’ and engineers’ experience working with and on Cardano, while it appears this list was initially focused on “innovations” and therefore research-backed work-streams. And I felt the need to write this reply in my own, personal, capacity. Let me repeat that: the views expressed in this text are my own, not ones from any organisation I am affiliated with.

In the past 4 years I have been working with an engineering/architecture hat nearly exclusively on R&D projects, interacting closely with researchers on well-known projects like Hydra, Mithril, ALBA, Peras, Leios, and of course Cardano itself. I also helped framed the Software Readiness Level scale as it’s used today inside and outside IO, and contributed to improving Research/Engineering collaboration and R&D Process. I have some experience with academia and research projects from previous positions, having myself gone through the ordeal of PhD defence and graduation. And last but not least, I personally appreciate working with researchers as they are both extremely smart and nice fellows.

And this experience lead me to the realisation that the way Research, and by that I mean scientific research, is understood and framed within the Cardano community is mostly wrong and not conducive to both high quality research and high quality software. I mean here Gerald Weinberg’s definition of quality from his seminal book series “Quality Software management”: Value to someone.

I would like to dismiss a common misconception which is the belief that “engineers implement research papers”. Actually, there’s no way any particular research work can have a direct connection with software in production. Of course the ideas, principles, concepts, and sometimes techniques can be ported from one domain to the other, but it’s extremely rare there can exist a direct (and I am not even speaking of a machine checkable) link between the content of a research and a software artifact.

Four examples will serve to illustrate that point. although I could possibly provide more than a dozen:

1. In all Ouroboros papers, the nodes in the network are supposed to exchange whole chains within Δ, the bounded delay for semi-synchronous networks. In practice this is obviously impossible and this has lead to the need to design a complex network and consensus stack to emulate that theoretical behaviour in practice
2. The protocol presented (and proven) in the Hydra paper is not the one that’s actually been implemented because it was actually too fragile as is and required extensions that were presented in an appendix, but actually had flaws
3. The whole Mithril network is but an example application of Mithril TMS scheme which is described in a couple of paragraphs at the end of the paper. Even the core cryptographic techniques form but a limited part of the content of the paper which is mostly filled with the details needed to prove the protocol in the Composition Framework theory
4. One of the thorniest part of implementing Leios over Cardano, how to handle the ledger, is a single paragraph in the whole Leios paper.

Moreover, scientific research has its own logic, processes, venues, people, groups, organisations which I collectively group under the term institutions, which have very few in common with the institutions of a large scale OSS development community, whose institutions have very few in common with ones that are needed for a financial market, etc.

At the core of (modern) scientific research lies the peer-reviewed publication process, and the goal of a researcher or a research team is to publish as many papers as possible. And not all publication channels are equal, some are more prestigious than others, with internal reports and conference workshops at the bottom, and renowned scientific journals like Theoretical Computer Science for CS or Nature for physics/biology at the top. In the academic world, whether publicly or privately founded, your career is based on your and your team’s publication stream, sometimes abstracted away in coarse grained metrics like H-Index.

But crucially, whether or not a paper is published depends on the peers review, a lengthy and time consuming process whereby experts read your paper, analyse it, possibly discuss and suggest amendments, ultimately deciding through some collective process whether or not the paper is deemed good for publication. And the key criteria underpinning this decision process is novelty: Does your paper proves something new? Does it introduce a new technique, a new algorithm? Does it remove limitations from existing ones? Does it use a new proof technique for an old problem?

Of course, the paper has also to be sound, reasonably well written, without any blatant error, and with a thorough coverage of previous literature on the particular topic it addresses. In theory, reviewers are supposed to guarantee the proofs are correct and while this might be the case in the most spectacular cases, eg. when someone publishes a proof of a previously known hard to prove conjecture in maths, chances are the poor reviewer who are themselves busy publishing, overseeing PhDs and post-docs, writing grant applications, won’t spend the time to rework through the excruciating details of the proof.

One consequence of this state of affair is that researchers have a poor incentive, and very few time, to work on not-so-novel issues and problems that would not interest their peers from the fields they are publishing in. Practical implementation details, usability issues, product-market fit, industrialisation, or practical performance and efficiency, are some of the details that are left for the reader to fill.

Another consequence is that scientific problems that can fill a life-time are quite rare, which means that once a problem has given rise to a more or less long streak of publications, it’s not hot anymore and researchers have to move on, similar to hunter-gatherer nomad tribes in deserts who need to move to a new spot once the resources of the current one are depleted.

These institutions and their sociological consequences have been studied extensively and can lead to all kind of perverse effects like cronyism, clientelism, entre-soi, or even fraud as some recent examples from even prestigious journals have shown. Not going to that extreme, what I mean to demonstrate here is that researchers and research teams have very few incentive to solve concrete problems happening to concrete people because chances are high these problems are uninteresting from a research perspective. Or a solution to a well-known problem would need to be quite sophisticated and therefore probably quite complicated to be considered worthy of publishing.

A key logical consequence of this situation and the exponentially growing amount of articles published, is that each scientific field tends to become overtime highly specialised as researchers need to compensate for increased sophistication through narrowing of the topics and hyper-specialisation. This dynamic has been hilariously and accurately portrayed in David Lodge’s Small World.

Each topic being researched by an ever smaller group of people also means the pool of reviewers jointly narrows, potentially to the point where the whole process is somewhat incestuous as reviewers, writers, and conference organisers all know and depend on each other leading to intricate conflicts of interest. But crucially this means that the relevance of a particular line of research, or the quality of a publication in that area, can only be meaningfully assessed by few people.

For the particular case of Cardano the situation I depicted above which, to the best of my knowledge, faithfully represents actual scientific research, has the following specific consequences:

• Research cannot be narrowly focused on Cardano, it has to be very general and rooted into known and accepted scientific problems that go way beyond the problems of Cardano,
• As outlined by the OP, research requires time and depth, and it therefore cannot be tied to monthly or even annual budget constraints
• One never knows whether a particular research idea or program will be fruitful,
• Research programs and even more importantly research results can only be evaluated by experts and it is demagogical to think that anyone who can create a DRep address and post the needed deposit on-chain can do it
• Research results must be evaluated on their own scientific merits and not on how fast they can be implemented on-chain, given the fact there’s possibly no actual problem in want of the solution researched
• Research work cannot be planned, but it can be evaluated and its course corrected.

I do think however research is critically important for Cardano to survive in the long run, and I do think IO researchers are doing a great research job. But when it comes to funding and evaluation, we should be careful: termporalities, success criteria, institutional environment, work organisation in research are different and this means decision process has to be different.

I also do think we should not fall into the demagogical trap: very few people, and I do not include myself in this group, are able to evaluate the merits of each of the projects mentioned above on scientific grounds, and none are able to evaluate all of them.

This lead me to suggest that a sound and meaningful Research budgeting process for Cardano would look like the following:

• Appoint a research board, possibly not permanent, populated by independent experts with sufficient credentials
• Ensure research board members do not receive any funding from Cardano (but are being compensated generously for their work on the board)
• Provision a global envelope spanning several years, and let the board allocate funds to programs and projects based on formal proposal process
• Regularly (at least yearly, possibly twice a year), evaluate research “grants” based on actual results (eg. publications, possibly software). But crucially, while the evaluation will be lead by experts, a vote from community would validate (or not) the board’s decision base on facts and actual outcomes
• Evaluate the impact of research on Cardano globally and over a multi-year period.

In short: Trust but verify, separate powers, and only expect long term returns.