Ethereans Turn Agains ProgPoW – Trustnodes

Ethereans Turn Agains ProgPoW


“In the interest of progress can we agree on a 5 minutes timeout for ProgPoW discussions.” – Vitalik Buterin, Ethereum’s Founder. 

ProgPoW is suddenly all the rage again with Santiment data showing this is currently one of the most discussed topic among ethereans.

It opened yet another dev call where about 30 minutes was again spent on whether ethereum should change its algo through ProgPoW to remove asics, with Hudson Jameson, a coder who is sort of in charge of the Ethereum Foundation (EF) public relations, making quite an astonishing statement.

ProgPoW is “to appease the current miners who said they want this,” Jameson said. He then publicly said it was some bargaining chip for miners to accept the new issuance reduction to 2eth per block.

Who exactly had such authority to engage in such negotiations with miners, or even why miners were asked at all about the issuance reduction, is unclear.

“I recently went back and listened to all the Core Dev calls that I could find in which ProgPow was discussed (call 38, and every call between 45-57 – let me know if I missed any) to understand the process by which we got to this point,” a dev says, adding:

“On call 52 the Core Devs made a decision to tentatively go ahead with ProgPow after seemingly only talking to miners and the IfDefElse team about it. They may have talked to others too but I’m only going by what’s reported on the calls.”

These calls are usually very boring affairs where highly technical minute details are discussed. It is more of a work like meeting, so usually no one really listens to them.

Sometime there are important decisions to be made. When the issuance reduction was discussed, for example, it was well publicized before hand, so most paid attention.

That a decision would be made on ProgPoW on this call 52 was not well publicized before hand, so really no one was paying attention. Except for Coindesk. For some reason one of their journalists decided to sit through this specific call with an article then stating a tentative decision to implement ProgPoW had been made.

That reopened the debate again, a debate that was closed during summer last year when Buterin effectively said ProgPoW wasn’t needed because Proof of Stake (PoS).

When the debate was re-opened, a vote was held where effectively 100% of mining pool operators that expressed a view voted in favor of ProgPoW. Precisely 100%, without any dissent whatever.

That obviously raised the question of where are these asics? If all miners want to change the algorithm to kick out asics, then presumably all current miners are GPU miners.

Ethereum’s current algo has a design that sort of requires the use of many GPU like functions. You can game it a bit. You can use so-called pills. You can use FPGAs or so called asics.

Bitmain launched an asics last year, but the initial reaction of miners was one of disappointment. Just 20% efficiency, a bit more than the pill.

There’s suggestions a Bitmain competitor, Linzhi, might launch more efficient eth asics. Linzhi however has said they’ll be able to do the same with ProgPoW.

Just what is marketing and what is fact remains to be seen, but what happened at Monero does reveal something important.

It can well be the case that as soon as the algo is changed, asics for the new algo are launched. After Monero changed its algo, it became 80% dominated with asics again after just a few weeks.

That ethereum still appears to not have any significant asics close to four years on and with such a huge market cap, is quite an achievement.

While all this is being discussed, however, Jameson now says they have made a decision and it is not tentative anymore.

“I am specifically saying that we have decided to implement ProgPoW into a future Ethereum fork and it will either be Istanbul or a fork that is yet to be decided on,” Jameson said.

Unlike easily sybilable reddit accounts, these are all real people and in many cases leading figures in ethereum. There are others too, including probably Buterin who has very subtly made it fairly clear that he’s against ProgPoW.

There are many reasons, but there’s now a new one. An apparent either developer of ProgPoW or someone very familiar with it, says:

“At current figures (ie. 7450000 blocks) the time spent by a node to validate block sealing nonces for all blocks is roughly 4 hours. Assuming a ProgPoW nonce requires 1.5x the time for the same validation we’d get at most 6 hours for the very same task. This is the only ‘overload’ the change of the algo could bring.

Reality is to sync a full node from scratch you need in average several days if not weeks so the marginal increment in PoW validation is less than negligible.”

So apparently ProgPoW makes block validation harder. Not by that much obviously, but some say it would increase uncle rates and thus inflation.

Devs have been working very hard to shave a millisecond here, a minute there, an hour in syncing, with every little bit counting.

Obviously if there is some very good reason then you can say fine, 10ms or an hour and a half isn’t that much, but you would need a very good reason.

With ProgPoW the only reason given is asics bad. Yes, arguably GPUs are bad too. Far better just CPU although even there you’d get CPU mining farms that make your laptop useless anyway.

Problem is ProgPoW is not an asics proof algo. It’s merely “resistant.” Just how much resistant, no one knows. So is ethereum then to spent time every six months forking off the algo with devs, the ecosystem and all the rest spending much time on this instead of pruning, state rent, lowering sync time, sharding, PoS, and whatever else devs come up with that actually improves the network.

If ethereum is not going to do this every six months, then why do it at all? Why deal with accusations that are already being made of bailouts and lobbyists and all the rest. Why suffer the potential embarrassment of Linzhi or someone else launching a ProgPoW asics maybe weeks after the algo is changed?

Does ethereum in that case then drop everything and fork again, delaying development further when there’s a ton of useful work to do?

That’s before bitcoiners even catch on. Never stop your opponent while he is making a mistake as they say. If this goes through, then you can be sure we’ll all hear what they say.

Not that it necessarily matters, just a bit of popcorn at worse, but it does raise questions about PoS if some ethereum devs are now negotiating with miners and exchanging bargaining chips.

In more familiar terms, special interests lobbying politicians shouldn’t have any say in ethereum’s development.

There is however a question of how do you make such decisions in a way that doesn’t give special interests an opening.

So far we have seen about three big decisions. The DAO fork had 97% of ethereans voting in favor. The Parity multi sig had again about 80%-90% against. The issuance reduction to 2 or 1 had some 97% in favor and about 60%-70% was in favor of reducing it to 1. The latter was ignored for some reason so as it happened price reduced to effectively 1eth per block.

Here, 95% have voted in favor of ProgPoW, but at least 1 million of the eth – 33% – and perhaps more come from one entity.

Another factor to consider is just how many “real” people are in favor or against. Here we’d say probably 50-50 at best and perhaps a majority against.

Ethereum has never had a change where there hasn’t been at least the appearance of overall agreement. Obviously in any decision you’d expect some dissent, with where exactly you draw the line being a hard question.

You can generally see however what the real people think and here it appears it is significantly above the level of mere dissent.

Obviously there are some in favor, but so many dapp founders, developers, and protocol developers being against isn’t something we have seen when previous contentious changes have been made.


Comments (2)

  1. A few milliseconds would not affect the uncle rates in any meaningful way. There are a lot of misconceptions like this one amomg the loud anti-ProgPow group. Another misconception is that it would suddenly take more time to do a full sync while the fact is, the few millisecond/block overhead will only apply to blocks after the fork, and before the POS transition, which depending on the time between forks could end up being less than an hour difference after a couple of years.

Leave a Reply

Your email address will not be published.

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>