[FRC] - Difficulty Adjustment fork SUCCESSFUL!

[Jorgeminator]

The fork has been moved forward to block #28336! Please read the OP.

[galambo]

Here's an example of how the filter works. This was produced in the process of testing the filter.

Have you ever seen a cryptocoin which achieves an appropriate difficulty in one day, starting from diff of 1?

That's what Freicoin does now.

[mhill12]

The change in the OP is slightly confusing, if I updated my local client to 0.0.2-2 a couple of days ago, do I need to update again?

The download link points to a copy of 0.0.2-2, so it looks like nothing has changed except when an upgrade to 0.0.2-2 is mandatory.

[galambo]

How long it would take for difficulty to retarget from 1 to 1+ million or the opposite case?

Bitcoin Megastore, I'm getting around 10 days but there is no telling how that would play out without really testing it.

[maaku]

The change in the OP is slightly confusing, if I updated my local client to 0.0.2-2 a couple of days ago, do I need to update again?

The download link points to a copy of 0.0.2-2, so it looks like nothing has changed except when an upgrade to 0.0.2-2 is mandatory.

It is confusing and for that I apologize. We took the exact same source code and changed the cutoff block to earlier (block #28336). The new binaries replaced the old binaries at the same URL. So download the binary again, and install it again overwriting the old files.

[Tobius]

The change in the OP is slightly confusing, if I updated my local client to 0.0.2-2 a couple of days ago, do I need to update again?

The download link points to a copy of 0.0.2-2, so it looks like nothing has changed except when an upgrade to 0.0.2-2 is mandatory.

It is confusing and for that I apologize. We took the exact same source code and changed the cutoff block to earlier (block #28336). The new binaries replaced the old binaries at the same URL. So download the binary again, and install it again overwriting the old files.

Lol, that's bound to cause some confusion

Hopefully it all turns out for the best

[maaku]

How long it would take for difficulty to retarget from 1 to 1+ million or the opposite case?

Bitcoin Megastore, I'm getting around 10 days but there is no telling how that would play out without really testing it.

Every 9 blocks it can adjust as much as +/- 1.055x. That's approximately a doubling (or halving) every 117 blocks, or a factor of +/- 10x every 387 blocks. So 1 to 1e6 or vice versa would take about 2322 blocks, plus another day/144 blocks or so for it to stabilize (the filter is slightly under-damped, so it overshoots and then corrects; you can see that in the real results above). How that works out in real time is a different story, because obviously 1 block != 10min for most of the adjustment period. From 1 to 1e9 about 10 days sounds about right and I trust galambo to have calculated that correctly. In the other direction it'll take quite a bit longer, as at first ~19 years (!) will pass between blocks. Hard-fork is your only real option in that case.

[Smiley_01]

I stumbled over the version number problem already with your last software change.
Why did you again not change the version number ?

Now, after installation/restart nobody can tell for sure if really the "newest" version
is running with the even earlier hard fork block ?
What if that happens with some of the bigger FRC pools ?

You test the new difficulty algorithm for days but don't take your time (about 1 minute?)
to change the version number....

Really, I want this coin to succeed, but I can't follow your reasoning here.

Thanks for your time & consideration.

[mhill12]

Version numbering is especially important whenever a change causes a hard fork. With two versions of 2-2 out, there exists a distinct possibility of an unexpected fork because of clients disagreeing about when the new difficulty calculations take effect. This new version should really be called 2-3 to avoid that confusion.

[galambo]

How long it would take for difficulty to retarget from 1 to 1+ million or the opposite case?

Bitcoin Megastore, I'm getting around 10 days but there is no telling how that would play out without really testing it.

Every 9 blocks it can adjust as much as +/- 1.055x. That's approximately a doubling (or halving) every 117 blocks, or a factor of +/- 10x every 387 blocks. So 1 to 1e6 or vice versa would take about 2322 blocks, plus another day/144 blocks or so for it to stabilize (the filter is slightly under-damped, so it overshoots and then corrects; you can see that in the real results above). How that works out in real time is a different story, because obviously 1 block != 10min for most of the adjustment period. From 1 to 1e9 about 10 days sounds about right and I trust galambo to have calculated that correctly. In the other direction it'll take quite a bit longer, as at first ~19 years (!) will pass between blocks. Hard-fork is your only real option in that case.

Thanks for explaining this in detail.

[coinerd]

I came in here to see how close the fork was and read this...

It is confusing and for that I apologize. We took the exact same source code and changed the cutoff block to earlier (block #28336). The new binaries replaced the old binaries at the same URL. So download the binary again, and install it again overwriting the old files.

Seriously?

You would be in some deep **** if you did that to financial software in any regulated market.

The support staff on your own team would kill you, if it was  a heavily used/supported tool in any market.

I realize that in this particular situation it's probably not going to cause much harm.. Still, after years of QA and configuration management, that gives me a headache.

You should really version at this point, while you still have a tiny bit of time left before the fork. How much easier is it to say "make sure you are on v3" (for example) than to go through "Which version of v2 do you have? ...You don't know? .... Well download it again and see if you still have the problem..."

I mean, this is the exact reason that the ENTIRE GLOBAL SOFTWARE DEVELOPMENT INDUSTRY agrees and uses version numbers on everything.

When you publish a mandatory upgrade to v 0.0.2-2 how many people are going to check their software, and be relieved that they "already have" the right version?

You've already floated two versions with the same ID and differing functionality.  In order to stop that from becoming a major problem for your product, I suggest you float a new version so that there is one version BEFORE the upgrade, and one version AFTER the upgrade.

No one should have to type this out for you at all. 

[SaltySpitoon]

how is this stickied?

I tend to sticky threads such as these when people need to know about a mandatory update by a certain time. It will be unstickied after block 28476. By the way, what block is FRC on now?

[maaku]

#28303. And the hard-fork is now scheduled for #28336.

[zero3112]

#28303. And the hard-fork is now scheduled for #28336.

How much longer before this is over?

[maaku]

Monday at the current hash rate.

[crazyearner]

remove

[crazyearner]

remove

[mhill12]

Well the more people who jump back on to FTC the quicker it will go down so more people is needed.


COME ON MINERS MOVE TO FTC FOR A DAY OR SO PUSH THE DIFFICULTY BACK DOWN YOU KNOW YOU WANT TO

This the Freicoin thread....

Freicoin has its hard fork at 28336.

[crazyearner]

Well the more people who jump back on to FTC the quicker it will go down so more people is needed.


COME ON MINERS MOVE TO FTC FOR A DAY OR SO PUSH THE DIFFICULTY BACK DOWN YOU KNOW YOU WANT TO

This the Freicoin thread....

Freicoin has its hard fork at 28336.

FFS sorry for that lol got 30 tabs open thought me in FTC lol

[mhill12]

It happens, although FRC really needs some people to mine until the difficulty adjusts. Network hash rate is down to ~100 GH/s. I should not be over 1% of the network hash rate with my 1.5 GH/s.