The Fight for On-Blockchain Bitcoin Scaling Soldiers On

A web conference this week found a notable selection of industry thought leaders exploring lesser-publicized challenges and solutions for bitcoin scaling.

Following a longer conference in June, Tuesday's "encore edition" of OnChain Scaling saw presentations from Bloq co-founder Jeff Garzik; Cornell professor Emin Gün Sirer; and Berger Singerman counsel Andrew Hinkes. A byproduct of the long-running block size debate, the event has emerged as a way to highlight proposed technical changes designed to be made to the bitcoin blockchain directly, rather than through a top-layer solution.

Though a topic of controversy, the latter path to expanding the network has proved the preferred choice for Bitcoin Core, the network’s largely volunteer developer group. To date, these members of the development community have largely promoted and prioritized efforts such as sidechains and the Lightning Network, which would move transactions off of the main blockchain to achieve greater transaction capacity.

In this light, it was perhaps Garzik who criticized Bitcoin Core's approach most strongly, arguing it goes so far as to be undemocratic. By not proposing changes to network participants in a way that allows them to select from two versions of the blockchain, he asserted that Bitcoin Core has become a central authority for decision-making.

Garzik told the audience:

Garzik's statements give voice to some in the bitcoin community who argue that the in-progress top-level networks might not work as well as planned, and that on-chain solutions, while more difficult to engineer, are equally worthy of exploration.

While on-chain scaling advocates are perhaps best-known for advocating for an increase to the block size, the three talks offered diverse views on bitcoin’s challenges, touching on alternative bitcoin protocols and the legality of ethereum's recent hard fork.

Overall, Garzik's presentation cautioned that the community should not grow too comfortable with the current favored way of making changes to the bitcoin network.

Whether to help scale or make some other upgrade to the network, there are two ways that developers can change bitcoin’s rules: hard forks and soft forks. Garzik argued that soft forks are not necessarily a superior alternative to hard forks, as many developers have argued.

To make his point, Garzik pointed to Segregated Witness, a popular scaling solution that's currently being rolled out in bitcoin. He agreed that SegWit is a useful change, but argued that it’s perhaps overly complex, and that it should be added in a way that bitcoin users have a way to voice their opinion.

As an example, he noted that SegWit would introduce a more complex fee structure than currently in place, a development that wallet developers like BitGo's Jameson Lopp acknowledged in the talk's question-and-answer session.

Garzik went as far as to argue a goal of the community should be to completely remove developers from the decision-making process to deal with this issue.

"Decisions made by humans should be replaced by algorithms as soon as possible," he said.

In another segment, Cornell professor Emin Gün Sirer outlined his project Bitcoin-NG, an alternative version bitcoin protocol he proposed in 2015, and many of Sirer’s comments were directed towards the aforementioned block size debate.

Sirer called Bitcoin-NG the "third way" and "a path out of the morass" in the debate, which has split the community between those who want to see an increase in the block size and those who want to scale transactions via other means.

He further offered advice on how he believes developers could best to determine whether to incorporate a change to bitcoin, scaling-related or otherwise. (How best to decide whether a change is good or bad has been in question the block size debate).

Sirer argued that an emphasis on science was missing from community discussions, and pointed to Cornell's cryptocurrency research arm's effort to test alternative protocols on a copy of the bitcoin network that’s low-cost to maintain.

"We cannot rely on gut feelings and vague concerns," he said.

The noted academic suggested that Cornell's testbed platform, 'Miniature World', could come to advance discussions, as it uses 1,000 nodes to test theoretical advancements.

Sirer said:

Sirer argued that this way of testing code changes is more efficient, realistic and easier, as it removes costs associated with mining, uses copies of real-world data and doesn't require potentially damaging test features to be implemented on an active blockchain.

Plastic chains image via Shutterstock