...The holier-than-thou approach of the team to an outpouring of disappointment from their early supporters and the toxicity that breeds has killed what I used to love coming to this community for....
...
We're weighing 2 very important things at the moment: Network security, stability, and reliability and what the community is saying. We have some (private) very important information/data on our end that is heavily contributing to our decision to focus where we are.
While I can understand the argument for continuing with core development - especially to address these nebulous "security" issues you keep alluding too - I also wish to invoke the apocryphal quote usually attributed to John Maynard Keynes* - "When the facts change, I change my mind; what do you do?"
More specifically, the fact that changed here is the network hashrate, which has gone from 5MH/s to as much as 100MH/s. Contrast that with ZenCash, for example, which has a network hashrate of 75-80MH/s and a market cap of $131M (all $ in USD), so is worth about $1.75 per MH/s. The situation is even more stark with Monero, which now has a hashrate of 530MH/s and a market cap of $4.5B, so is worth about $8.50 per MH/s. Extrapolating these two examples to DERO would result in a market cap of somewhere between $87.5M ($1.75 * 50Mh/s) to $850M ($8.50 * 100MH/s). Does even the lower valuation seem at all realistic given the recent market conditions and the near total absence of any actual use for DERO (at the moment)? Not to me, it doesn't.
I rather suspect the pitchforks and cudgels would be brought out if DERO's price had continued its downward trajectory, but fortunately for all it seems to have not only stopped declining, but reversed smartly to the upside, and so the community is merely upset, rather than furious. But again, the facts have changed, --Serena-- and CaptDero; will you change your minds, or stubbornly continue on the same path?
* - but most likely first said by another economist, Paul Samuelson
I think that you and I are generally of a like mind, not just where DERO is concerned but also regarding other projects in which we have a mutual interest, but I don't agree with you 100% on this one.
In real life, something that I'm regularly faced with is stress testing applications, the web server stacks that serve them, and the hardware that it all runs on. One of the most difficult things to come up with is "actual usage" tests (rather than something as simplistic as ApacheBench or similar) that simulate real world usage patterns. Part of the reason that it's so difficult is because there's nothing even close to a standard usage pattern when hosting mixed-application websites belonging to hundreds of different people and with wildly varying traffic patterns, application configurations, resource requirements etc.
One can easily see where the emergence of ASICs and the sudden surge in attacks and the network hash rate has dropped a rather useful "real world" stress test right into the DERO developers' laps. Consider what other projects have done. ITNS was attacked and essentially went offline while a hardfork was implemented and the chain rolled back. XVG was attacked and basically exposed the absolute incompetence of the developers who had no idea how to respond. There are other examples I'm sure, but those are the two that I'm most familiar with.
Imagine that ITNS was more mature and in heavy worldwide usage. Can you picture the sheer chaos caused by a hardfork and roll back of the chain? Even if the usage was related to its primary function (a mechanism for VPN services) and not more widespread commerce, that would mean that payments to providers would essentially be undone for the entire duration of the roll back, and that the VPN services that people had come to rely on were not available to them. Would ITNS be trusted again in such a scenario, even after the fix had been implemented? Maybe. Maybe not. There would be an awful lot of VPN service providers who would be mightily pissed, and rightly so.
DERO, as you rightly said, doesn't currently have an actual use case. There are no end-users to inconvenience, no commercial transactions to be interrupted or rolled back. It's a new blockchain technology in its infancy, and the only people being inconvenienced are GPU and CPU miners. Does the current surge in the network hash rate suck for them in the short term? Sure. Though it's really not that hard to mine something else and exchange it for DERO at a rate far higher than can be obtained by just mining DERO. Example: I can mine less than 3 DERO per day with my measly 10KH/s, but I can mine XMR and buy more than 20 DERO per day. That's not far off what I was mining before the ASIC invasion.
The DERO developers however are getting that real life stress test that I talked about earlier, and they're getting it for free. The network has been attacked, the attacks have failed. The hash rate is 10x what it was pre-ASICs and yet everything continues to run smoothly. Heck, just look at the number of Alpha releases that came one after the other in response to the attacks and network surge. The devs were clearly learning, and learning fast, not just from what was happening to DERO but to other projects too. You simply cannot buy that kind of data about how a blockchain will respond to suddenly changing conditions and / or attacks of various amplitudes.
With that in mind, I'm not at all convinced that an immediate hard fork to ward off the evil ASICs is (or was) necessarily in the best long term interests of the DERO project, its supporters, and its investors. Nothing is learned that way. And when new ASICs come out that can mine v7, Lite, Heavy etc the projects that swept network / security issues under the rug with a simple algo change will find themselves in the same situation again.
The more I think about it, the more I'm convinced that the DERO devs should be applauded, not castigated, for taking this free-of-charge real world stress test, patiently learning lessons from it and applying them to the underlying blockchain technology, and THEN eventually making a switch to a different algo once they've learned all that they can. In doing so, they're showing the kind of foresight and discipline that we should all want in the devs of a project that we hope will flourish into something massive.
Short term pain for long term gains. I can live with that.
P.S. Wanted to add some quick math since I know you love it as much as I do. Pre-ASICs my 10KH/s was mining ~25-30 DERO per day, and by mining XMR then exchanging it for DERO I can currently buy ~20 DERO per day. Worst case scenario, it's costing me 10 DERO per day. Let's say that DERO doesn't implement a new algo for 2 months - 60 days. The total cost to me is 600 DERO, or at the current price about $540. Will I trade that for all of the lessons that are being learned and put into practice by the devs? Every day, and twice on Sundays. If DERO takes off as we all hope it will, it's well worth that to me to avoid the pains that the ETH network experienced.