Hacker News Reader

Comment by out_of_protocol | Hacker News Reader

Why segwit is not good:

* Realistically, can not be activated (95% support level required, way more than achievable). Means we'll stuck with current 1MB forever waiting for segwit

* Upon activation it does not add any benefits. Real increase in throughput will be achieved later, then all major players updated their client code, started crafting segwit transactions and moved coins to new segwit addresses. Means it delays throughput increase even more

* After full adoption, throughput will be increased from 1MB/10min up to ~1.7-2MB/10min while worst-case scenario (useless, specially crafted transactions) is 4MB - miners should be prepared for it. Means "2MB forever" and "hard to increase block size because miners should be prepared for 4x load"

There are other points as well, i'll stop here. Simple block size increase don't have such cons

View on HN

> Realistically, can not be activated (95% support level required, way more than achievable). Means we'll stuck with current 1MB forever waiting for segwit

This is not an argument against segwit, but against the method of it's activation. The rationale behind this high activation treshold is well summarized by greg maxwell [1]

> Upon activation it does not add any benefits. Real increase in throughput will be achieved later, then all major players updated their client code, started crafting segwit transactions and moved coins to new segwit addresses. Means it delays throughput increase even more

If only merchants or services who need to do lots transactions enable segwit, everyone benefits from the compressed output as there is more space on the chain for non-segwit wallets even if they don't upgrade or are aware of what segwit is

> After full adoption, throughput will be increased from 1MB/10min up to ~1.7-2MB/10min while worst-case scenario (useless, specially crafted transactions) is 4MB - miners should be prepared for it. Means "2MB forever" and "hard to increase block size because miners should be prepared for 4x load"

Miners should be prepared to give the service users want and pay for with fees. "Specially crafted transactions" included.

[1] https://lists.linuxfoundation.org/pipermail/bitcoin-dev/2017...

* It can be activated through UASF or through BIP149 which supposedly (big supposedly here) has been agreed to by a majority of miners

* It fixes transaction malleability and quadratic hashing problem. Transaction malleability is definitely a benefit upon activation. Oh... and it's also a block size increase so there is that benefit. Oh... and lightning network is basically ready. Oh... and rootstock. If the network NEEDS NEEDS NEEDS more capacity so badly, then throughput increases should happen very very quickly, because why would people wait to upgrade if the need is so incredibly dire.

* Don't even know what you're arguing here but "2MB forever" is not grounded in any sort of reality except one that people have made up.

Reply on news.ycombinator.com