Hey Shelby - I'll circle back to your previous post after this (that one requires more thought). First, I need to say that I don't 'speak for BU'. I'm just a guy that thinks their vision is correct. Most intelligent BU discussion does not happen on the more core-centric venues. I'm just trying to dispel a lot of the falsehoods told here about BU and the intent of its proponents.
To all Bitcoin Unlimited supporters, I am not against larger blocks.
Good. That's a starting point for a dialogue.
My thought is reformulate your protocol to have some where between 2 - 8MB block size limit increasing with the
Nielsen Law of bandwidth increase (slower exponential growth than Moore's law).
But BU
has a block limit. The difference here is that it is adjusted in an ongoing manner by a community process. If 2 or 8MB is the proper value at one point in time, then that will be the limit. If we go with some other fixed value, we'll just be fighting this battle again in no time, with all the wasted friction, animosity, and market turmoil that it entails.
I realize there are skeptics that scoff at the emergent consensus idea, but there is science behind this concept. Granted, its application to this field is novel. At least as surfaced in this manner. But in the end, each party has the ability to change the blocksize they'll accept anyhow, by a localized change to source and recompiling. We are just making it easier for the user to make this change on the fly.
From my perspective, no. You are not going to convince me to abandon the scalable blocksize vision. With the caveat that I'm still thinking about your previous post.
I detect no deviation in the BU community at large from this principle either. Especially as -- after over a year of hard effort -- the tide seems to be turning in our direction.
Also you must adopt some of the bug fixes in SegWit, such as the new opcode which fixes malleability and enables Lightning Networks.
By and large, there is no effort against adoption of segwit (the targeted feature in isolation - not the bundle of cruft which is The SegWit Omnibus Changeset) within the BU community. We recognize that malleability is something that is worthy of addressing. However, it is really a tertiary concern. For all the heat and light, I have yet to be shown that malleability has actually caused any systemic failure. Or significant value loss. The restriction of ~250,000 transactions per day, on the other hand, is a systemic problem that needs fixing RFN.
So yes, let us fix malleability. Once we are past this tps barrier. If my read of the BU community is correct, segwit (likely without the centrally-planned magic number signature discount) is a contender for this fix. There is a conversation comparing and contrasting with (e.g.) FlexTrans that first needs to happen though.
But do not adopt all the rest of the complexity of SegWit, including do not adopt the ability to softfork version changes as that hands too much power to Core.
Agreed. Note that the feature of which you speak is _not_ the actual segwit feature, but what I refer to above as part of 'the bundle of cruft which is The SegWit Omnibus Changeset'. I don't believe that the majority of the BU community supports this either.
Go do that, then maybe you can win. Because likely users will end up choosing to support your protocol because their transactions are not being delayed. But you need to build confidence in your competence to lead.
You can possibly win, if you get competent people working with you.
You are apparently competing against banksters who are funding the Core developers (paying their large salaries and incentives). Perhaps that may explain why you ostensibly can't attract the most competent developers?
(I've taken the liberty reorder a couple of your paragraphs)
For the most part, the BU community realizes that our dev workforce is not comparable to core's. We have no illusions on this point. The fact that the most significant core devs are highly paid is likely a factor, but probably not the only factor. Personally, I think a very large factor is that devs naturally want to work on the market-leading implementation of a given idea. Up until recently, that has unambiguously meant core. However, from my perspective, there has been a recent swing in the thought-space towards BU. If indeed BU can wrest the market share title, I would expect a commensurate movement of devs from core to BU. Of course there will be attrition, but we don't need many to double or quadruple our dev bandwidth.
(One other factor would be untruths repeated far and wide about features, capabilities, and even motivations of BU - some intentional and some merely repetition of falsehoods. But getting into shouting matches over this would be unproductive)
Yes, there have been bugs. Fortunately, none have caused persistent or systemic issues. And likely, more will be found. (Of course, more will likely be found within core as well). But a mediocre implementation of the right design is infinitely more valuable than a nearly-perfect implementation of the wrong design. For the former will asymptotically approach the ideal, while the latter will always be targeting off in the wrong direction. We expect BU will be that goodness.
But you need to rein in these talking heads who are not sufficiently competent, i.e. Wu and Ver. They can talk when they cite competent developers. Those guys need to understand the limits of their role and competencies.
Look. This narrative really ought to stop.
Neither Ver nor Wu have any official relation to the BU community. Are they visible BU community members speaking their minds? Yes - but only in their personal capacities. The idea that they are somehow official thought leaders in the BU community is merely a false redditism. Neither of them were responsible for brainstorming nor initiating BU, and neither of them participate regularly in discussions in the primary BU venue.
I am not against what you want in terms of fixing the block size problem. I am just being pragmatic. Find a confidence building and realistic way to win. I don't like being on the losing side.
I highly advise that you all back down from your current ultimatum and reformulate and get the token holders on your side before you proceed.
Backing down is unlikely at this point. I'm certainly not.
The majority of Bitcoiners are likely unsure or ambivalent of a choice between core and BU. There is a hard core minority and a hard BU minority. I don't know the proportions, but BU seems ascendant at this moment.
No discussion of sentiment would be complete without mention of the intentional manipulation of the dialogue. While this has recently lifted to a large degree, it is undeniable that the discussion in this and other venues has been steered to marginalize BU. Accordingly, many BU supporters have left these controlled venues for others. As a consequence, core supporters -- whether out of ignorance or out of malice -- have had free reign to poison the discussion by parroting untruths re: BU. This has two consequences: 1) the actual number of BU supporters is likely larger than you perceive; and 2) this number is likely to grow as misconceptions are corrected.
You really need a more capable leader than Wu or Ver. But don't enlist me, because I am too busy on my own "shitcoin" project and I am not interested in investing my effort in Satoshi's PoW. It is good to have those guys on your side, but not as the outspoken de facto leaders, because they have by now proven they are rash, irrational, and somewhat technologically incompetent.
Again. Neither Ver nor Wu are in any way central to the BU project. They are merely passionate supporters much like myself. They just have bigger megaphones than I due to their stature in other Bitcoin areas.
And yes - I would like to see you devote more time to your alt. While I am skeptical, not having seen your paper -- and despite our occasional inability to see eye-to-eye in the past -- I respect your intellect. Someone will make the next breakthrough in cryptocurrencies. It might be you? I have no way to judge currently, but I would not find it inconceivable.