Pages:
Author

Topic: Current SegWit code does not fix O(n^2) - page 2. (Read 1346 times)

legendary
Activity: 2674
Merit: 3000
Terminated.
June 28, 2016, 11:27:33 PM
#5
Wow all the hype and it's not going to help clean up the mem pool right away, I'm I getting that right?
No, you're not getting this right. This has nothing to do with "help clean up the mempool".

Like the title says. The current iteration of The SegWit Omnibus Changeset does not fix the O(n^2) hashing problem. At least according to Peter Todd:

Quote
We haven’t actually fixed the O(n²) signature hashing problem yet, although we’re fairly confident that we can, and there’s a open pull-req implementing the cache that we need.
While I must admit that I was unaware of it at first, I don't see this as something problematic especially if you consider the last part. I've read some parts of that review last night and it seems very well done.

Might be worth a discussion, donchathink?
Maybe. However, don't be surprise if you see a lot of 'unusual' accounts joining in to bash Segwit.

I'm kidding. I just thought that with all the press in the BTC space of its implementation going live, we would actualy see it working and cleaning up he pool. That's all, just disappointing it's not ganna work out that way right now.
Again, you have no idea what you're talking about. There's no reason for it "not to work out that way right now".
legendary
Activity: 1148
Merit: 1000
June 28, 2016, 11:15:16 PM
#4
1 THING:  DONT RUSH THEM.

Whatever happens, dont rush the devs. They are working hard and they dont need pressure on their heads.

If you rush them, they might fuck it up, and we dont want that to happen.

Or do we?! Wink cheap coins.

I'm kidding. I just thought that with all the press in the BTC space of its implementation going live, we would actualy see it working and cleaning up he pool. That's all, just disappointing it's not ganna work out that way right now.
hero member
Activity: 854
Merit: 1009
JAYCE DESIGNS - http://bit.ly/1tmgIwK
June 28, 2016, 11:11:36 PM
#3
1 THING:  DONT RUSH THEM.

Whatever happens, dont rush the devs. They are working hard and they dont need pressure on their heads.

If you rush them, they might fuck it up, and we dont want that to happen. The network can easily work well 5-6 more months, so no need to rush things.

They should take their time, think ,research, develop, and test many times the proposals. Then we can implement it.
legendary
Activity: 1148
Merit: 1000
June 28, 2016, 11:05:46 PM
#2
Wow all the hype and it's not going to help clean up the mem pool right away, I'm I getting that right?
legendary
Activity: 3080
Merit: 1688
lose: unfind ... loose: untight
June 28, 2016, 10:51:41 PM
#1
Like the title says. The current iteration of The SegWit Omnibus Changeset does not fix the O(n^2) hashing problem. At least according to Peter Todd:

Quote
We haven’t actually fixed the O(n²) signature hashing problem yet, although we’re fairly confident that we can, and there’s a open pull-req implementing the cache that we need.

- https://petertodd.org/2016/segwit-consensus-critical-code-review

Not necessarily an insurmountable problem. And I suppose PT might be... err... uninformed. However, it certainly puts some specious claims (e.g., April; e.g. safe scaling) into perspective. One wonders what other major claimed features of The Omnibus SegWit Changeset remain technical pauperism.

Might be worth a discussion, donchathink?
Pages:
Jump to: