Metropolis Pushed Off to January, Courts Criticism – Trustnodes

Metropolis Pushed Off to January, Courts Criticism


Ethereum abstract

Ethereum will see no upgrade this year with the issuance reduction pushed off to January or maybe February.

“The earliest we gone have it is January,” Hudson Jameson said on a devs’ call meeting livestreamed this Friday, with all in attendance being in consensus.

Devcon was one cited reason as many eth developers will be participating in ethereum’s biggest event of the year. While December had previously been ruled out because of the holidays, meaning many devs may not be around during the month.

Dimitry Khoklov, who was in charge of testing, said he hoped to finish it all by next week before Devcon. Afri Schoedon of Parity said he felt this was all being rushed. Martin Holst Swende of Geth said if this was going to be delayed, he was going to push for ProgPow, a proposed proof of work change for ethereum to get rid of asics.

Lane Rettig raised questions regarding project management. Certain information were hard to find, he said. Some of the devs may have a task list memorized, but there were things from the field of project management that could be incorporated.

Khoklov suggested there maybe should be six months of testing for the next hardfork, which they’re calling Istanbul. How that would synchronize with the Casper upgrade wasn’t discussed.

This upgrade was initially planned for 2017, with it called just Metropolis back then. They decided to split it into two, Metropolis Byzantium which went live in circa September 2017. Then the current Metropolis Constantinople which was meant to include the features that were not ready for Byzantium.

The then Hybrid Casper upgrade, however, was making progress with it launched on testnet in December 2017. So Constantinople was kind of shelved. Hybrid Casper then appeared imminent in spring, but a change of plans to prioritize sharding shelved that upgrade, with Constantinople revived.

Constantinople was sort of finished, but a decision had to be made during summer on whether to reduce issuance in light of the need to delay a coded exponential increase of difficulty designed to discard the Proof of Work chain so as to move to Proof of Stake.

If it was to be reduced, a further decision had to be made on whether to reduce it to 2eth per block or 1eth from the current 3eth.

The vast majority of eth holders voted for 1eth. Miners were invited to a dev call for the first time ever, they argued to keep it at 3. A compromise decision was made for 2eth.

That decision was made at the end of August. The change was very simple, devs said. They needed a month, however, to confirm the testnet block, a block that was set about two weeks ahead of the then end of September.

On a somewhat disorganized launch, the testnet found a consensus bug which appears to have been due to lack of communication with a Parity dev saying Geth was more involved with specifications and they may not communicate certain assumptions which might not be obvious to Parity.

Schoedon immediately called off the fork planned for November, saying it will now be in January prior to any public discussions with other devs. That’s while another dev from Parity stayed up late to fix the bug within hours. A new client was released within two days.

Khoklov was now planning to work hard and get all the testing done before Devcon so that he can relax and enjoy the nearly week long event. There would have then been another two-three weeks of testing, but end of November was a bit tight.

Realistically, mid-December might have been somewhat “easy,” with the Devs then free to enjoy Christmas, but some devs might have been on holiday even during the first two weeks, or they might have been other delays which may have pushed it much closer to Christmas. So they’re now planning it for January.

By planning, as far as this call is concerned, they said it won’t be before January. They haven’t quite decided when it will be, or when they will decide.

If the whole of December is written off, then they might decide a block number in January, but February or maybe even March might perhaps be more realistic.

Online discussion has been heated on the news. “Ethereum development is turning into federal government levels of slowness, bureaucratic delays, and extended holiday breaks,” the top voted comment says. While Schoedon says:

“Stop pointing fingers at Parity. Stop building up pressure on the core developers. We are trying to stick to high quality standards and you should appreciate that we rather delay a risky fork than just break the main network.

And no, we are not U.S. Congress, but a bunch of decentralized working engineer teams with remote working developers all around the globe and even though some appear to be available 24/7 most of us have family, kids, some even have a wife, and I even met someone who has hobbies two years ago.”

There is a question of accountability with all open source projects. And really we should have expected that all which can go wrong, will go wrong, as in some ways it has.

There have of course been many improvements and plenty has gone on in other aspects, but there appears to be different teams in eth which don’t seem to talk to each other very much.

There’s the sharding team, the casper team, the metropolis team, and not a team that kind of oversees the connections. Someone that knew what was going on in casper and sharding and could have said at the beginning of the year, rather than at the very end, that maybe they should be merged.

Or someone who knew what was going on with that merger and told the Metropolis team that the fork should perhaps be revived. The Metropolis devs, instead, appear to have learned of the fork at the same time as everyone else.

Likewise, those working on Hybrid Casper learned their code was to be discarded at the same time as the wider public.

So rather than bureaucracy, the problem might have been the opposite, lack of coordination. That may have led to time wasted, with code paused or thrown out when resources could have gone to far better uses.

Lessons learned, hopefully.



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>