Ethereum's world computer is in need of a magic touch.

As developers continue to clash over yet another controversial software update, some of ethereum's top minds are working together to conjure up a solution for what has been a troubling absence of late — community consensus.

Recent discussions concerning the return of lost funds — and a type of code fix that necessitates an "irregular state change," or platform-wide software revision — has led to internal conflict, with developers questioning their authority to make contentious changes while appealing to the public for opinions on the matter.

Strung up in the aftermath of the Parity fund freeze, a new developer collective is seeking to better organize such debates to achieve the kind of global consensus they believe the project requires to move forward.

Led by developer Greg Colvin and the Ethereum Foundation's Jamie Pitts, the Fellowship of Ethereum Magicians hopes to provide a structured working group where ethereum coders can coordinate in line with existing best practices for open-source development.

"We're talking about, 'Gee, you know, there's many, many millions of dollars just stranded out there for no good reason that we technically could fix, and should we?' And again, we don't have the forums to come to community consensus on these things," Colvin told CoinDesk.

Often occurring as a result of faulty code, some developers see the return of lost funds as an obligation, while others feel that such actions could be potentially criminal — a polarization that has led to bitter infighting.

"I think having that level of collegiality among the researchers and developers makes it easier for those conversations to take place and stay civil," Colvin said.

Worsening the state of the situation is that, in the case of decentralized protocols, any disagreement could lead to competing versions of the software – as occurred following the DAO hack of 2016, which led to the creation of a rivaling cryptocurrency code base called ethereum classic.

"I think the DAO is an example of making a big move without an adequate consensus," Colvin said, adding:

To Colvin's point, the recent dispute has revealed fault lines in the platform's development process overall.

Originally proposed to simplify the process for implementing fund returns, EIP 867 was criticized by some, with EIP editor Yoichi Hirai flatly refusing to even merge the proposal at first. Hirai's decision, along with the proposal itself, has pushed the community to rethink how much changes should be implemented — with some arguing that the process is too centralized.

"I don't want to be part of the ethereum community anymore if only one entity can singlehandedly block any proposal," Parity’s Afri Schoedon wrote on Twitter.

According to Colvin, such struggles hinge on how quickly the community has grown.

"The core developers initially were a pretty small group who all knew each other," Colvin said. While in these early stages, technical decisions could occur more easily, at this point he said, "It's a much larger group, it's spread all around the world."

Having first publicized the call for participation on reddit, the fellowship is set to begin with a workshop at the upcoming ethereum community conference, EthCC, next month in Paris. From there, Colvin hopes this will expand into a dedicated council by July.

And Colvin maintains that these in-person meetings can do a lot for resolving technical conflicts that can persist online for years.

"Sometimes you need to sit down in person and actually get to know somebody and establish a level of communication that wasn't there before," Colvin said.

In this way, the fellowship models its structure off the Internet Engineering Task Force, or IETF, an international collective of technicians devoted to the upkeep of the internet.

"I saw the IETF is probably the most relevant example of a success in that domain," Colvin said. Having "kept the internet running for many years," according to Colvin, the IEFT is built to cope with large numbers and does this by combining larger assemblies with smaller, more specialized groups.

"Each group does its thing, and it would be a rare nerd who would be interested in lots of these groups," Colvin continued.

Like the IEFT, the fellowship's governance process advocates what is called "rough consensus and running code," meaning that the dominant majority within a given discussion will be given precedence, dependent of course on its technical proficiency – the key criteria of judgement for any position held within the group.

Crucially for Colvin, the IEFT achieves this without any kind of corporate funding or any other sponsorship body that could in some way influence the activity of the collective.

Colvin summarized:

By combining an open process, an informal membership structure and an emphasis on technical responsibility, Colvin hopes that the fellowship can finally provide an adequate platform for the community to resolve its more sensitive topics.

Indeed, the question of lost funds isn’t the only technical crossroads ethereum is facing today — and arguably, it's a trivial discussion compared to the changes that are due to be implemented down the line.

"We're talking about moving from proof-of-work to proof-of-stake, when we're still designing proof-of-stake, when it's still not totally clear it will work. We're talking about moving into sharding, with two or three different designs in flux. We're talking really big changes to the protocol," Colvin said.

According to him, it's crucial that the wider development community, and not just the core developers, have a voice in these fundamental changes.

"Do we move to proof-of-stake? There's more than a small number of people who care about that and who know about that and have something to contribute," Colvin continued, concluding:

Greg Colvin via Status