ny2cafuse
Legendary
Offline
Activity: 1582
Merit: 1002
HODL for life.
|
|
December 30, 2014, 08:37:50 PM |
|
TL;DR: the change to 25% doesn't do what was expected and will only introduce chain splits. That's an assumption or has been tested with the simulator? We're discussing it now. -Fuse
|
Community > Devs
|
|
|
BioMike
Legendary
Offline
Activity: 1658
Merit: 1001
|
|
December 30, 2014, 08:50:39 PM |
|
How changing the max adjustment to 25% would cause a chainsplit, step by step: Say two miners (A and B) would mine a block at the same time and broadcast it to the network. Half the network gets block from A first, other half gets block from B first. The group that got block A will now perform an anti ddos check on the block from B, because B was later and therefore is considered to be a split. Lets assume that the digishield changed the difficulty over 25%. This means group A will completely ignore block B. In the meanwhile, group B has more miners and mines blocks B2 and B3; they now have the larger chain: "B>B2>B3". But since group A won't accept block B, they will never get this chain and won't resolve the split. The miners in group A won't either, they'll continue mining as if nothing ever happened. The same applies for group B: group B will never accept block A because it has > 25% difficulty change, so IF group A gets the longer chain (which would normally resolve the split), it won't be accepted by group B because group B will never accept block A. The normal max adjustment is 400%, which will allow even the most extreme diff changes by digi or dgw3.
How would such situation be resolved in non-digi algo's?
|
|
|
|
/GeertJohan
|
|
December 30, 2014, 08:51:41 PM |
|
TL;DR: the change to 25% doesn't do what was expected and will only introduce chain splits. That's an assumption or has been tested with the simulator? Neither, it's deducted by reading the sources.
|
|
|
|
/GeertJohan
|
|
December 30, 2014, 08:53:37 PM |
|
... bla bla bla block A block B blabla ....
How would such situation be resolved in non-digi algo's? No this isn't digi's fault. We simply shouldn't change ComputeMinWork's max adjustment from 400% to 25%. I don't think any coin ever did this, unless you have a very very extreme difficulty re-adjustment algorithm that actually could cause adjustments over 400%, in which case you want to change it upwards!
|
|
|
|
Frais
|
|
December 30, 2014, 09:02:13 PM Last edit: December 30, 2014, 09:15:54 PM by Frais |
|
TL;DR: the change to 25% doesn't do what was expected and will only introduce chain splits. That's an assumption or has been tested with the simulator? Neither, it's deducted by reading the sources. I'm getting confused, did the Criptoe team made some thinking errors?
|
|
|
|
veertje
Legendary
Offline
Activity: 952
Merit: 1000
|
|
December 30, 2014, 09:06:48 PM |
|
If testing can be done with Digishield in the simulator so soon, there where also discussions on changing parameters in DGW3 2 months ago orso that could do a lot if I am correct. If the simulator can really simulate small and big waves that soon, then take some time for good testing different parameters and blocktimes. Just my thoughts. 1 or 2 weeks more Clever, I can live with that.
You're making an excellent point here: yes we can try DGW3 with alternative settings in the simulator. We can even change formula's (not just the settings/parameters for the formula's). I think the 2,5 minute blocktime is of big influence as well. That does probably a lot. I wish you GJ and all involved in testing lots of succes in finding out the best solution. Thanks for updating. "But I'm not a coder..." Thats okay.. You might not be able to help with the more technical aspects of Guldencoin but that is just a very small part of what we're doing here. For instance: contact merchants, introduce your friends to Guldencoin (remember: they can get 100 free NLG! Who doesn't like free stuff!?) Or write an article about Guldecoin. Comment on large news websites on topics related to Guldencoin. etc. etc. Better even: think out of the box. The things I've just listed have been listed before, and although they are very very important, it's also good to have new and innovative usage and promotion of Guldencoin.. In the new year I will post way less here on the forums and do more real life approach of merchants.
|
|
|
|
Jero
|
|
December 30, 2014, 09:38:08 PM |
|
In the new year I will post way less here on the forums and do more real life approach of merchants.
Always online and doing great (support) jobs. Wish you good luck and succes for next year, approaching merchants is a real challenge. My respect you have!
|
|
|
|
nickelback65
Member
Offline
Activity: 85
Merit: 10
|
|
December 30, 2014, 10:22:49 PM |
|
Sending along some Guldencoins to the NLG tip jar. Keep up the great work.
|
|
|
|
ny2cafuse
Legendary
Offline
Activity: 1582
Merit: 1002
HODL for life.
|
|
December 30, 2014, 11:10:12 PM |
|
I'm getting confused, did the Criptoe team made some thinking errors?
Yes, there were errors in the code I submitted. There are 2 that I can discern from my talks with /GJ today. The first was the 25% limit that wasn't a 25% limit. It was a validation check that acted like a limit on our testnet by rejecting blocks on the confirming nodes. See /GJ's post about this, although all the a/b/c stuff totally lost me. The second part to that line of code was that if it was added to the network now, it would have cause the chain to not sync properly on a resync. Essentially the existing blocks higher than a 25% retarget would have been rejected. So it is my understanding that the results that were shown with the charts were correct because they were valid with a fresh chain on our testnet. However, they wouldn't have been valid on the main chain. Had we worked with the existing blockchain data, we would have identified the error immediately. My team missed it, and I sincerely apologize for it. I take full responsibility for these errors. I knew the code wasn't a final draft. That is why when Rijk asked me to post the git pull, I asked him to have a few pairs of eyes on the code before we did anything. It's also why, in light of the simulator timeline /GJ presented, I asked that we test against the simulator. /GJ is currently working on cleaning up the DIGI code. I have offered to reach out to the DIGI dev in the meantime. However, /GJ is confident that he will be able to provide a streamlined DIGI solution, so I am going to stand behind that decision. I'm going to take a step back from NLG for a little bit. I just need to refocus on what I'm doing. I'm not cashing out, and I'm not shutting down the pool. I will continue to sell physical coins and try to figure out ways to make NLG better outside of the blockchain. I just feel with the animosity over my involvement, and the movement towards finally fixing the algorithm problems, I don't need to be in the forefront for a while. I'll continue to follow the thread, but I probably won't push for radical changes like I have in the past. I'm just going to sit back and watch for a while. Cheers, Fuse
|
Community > Devs
|
|
|
Sharkzz1
Sr. Member
Offline
Activity: 880
Merit: 251
Think differently
|
|
December 30, 2014, 11:10:35 PM |
|
"But I'm not a coder..." Thats okay.. You might not be able to help with the more technical aspects of Guldencoin but that is just a very small part of what we're doing here. For instance: contact merchants, introduce your friends to Guldencoin (remember: they can get 100 free NLG! Who doesn't like free stuff!?) Or write an article about Guldecoin. Comment on large news websites on topics related to Guldencoin. etc. etc. Better even: think out of the box. The things I've just listed have been listed before, and although they are very very important, it's also good to have new and innovative usage and promotion of Guldencoin..
i don't know if the poster project from august is already done or not? if it's still not done then i could help with it. i think it's a great way to advertise for nlg. https://forum.guldencoin.com/index.php?topic=392.msg2844#msg2844someone that wants to help me with this project? (if it's not done yet) edit: sorry the topic is in dutch, it's just traveling to some train stations in big cities and posting some posters that's all
|
|
|
|
LTEX
Legendary
Offline
Activity: 1023
Merit: 1000
ltex.nl
|
|
December 30, 2014, 11:59:12 PM |
|
Wow! Here I am, practicing staying up late for New Years Eve tomorrow. I thought it would be boring, but hey, so much reading material! On a more serious note, what is happening right now is not a bad thing. Our beloved coin is getting to a phase where it's growing up, along with it's community. This can put up some serious challenges along it's path. Let's call it puberty.... The discussions that evolve around the Algo change are intensifying and this always stirs up emotions. Sometimes, especially when discussions aren't face-to-face, things get misinterpreted or twisted. This doesn't mean, however, they are no good! We sometimes need to cross lines to define new ones. To me, anyone involved in these discussions deserve my respect. I believe this because I haven't seen any participant that hasn't put in significant efforts in the (relatively short) past of this thread. It proves to me all opinions are defended by persons that have embraced NLG in their hearts and therefor feel they have to defend their stands. Just a couple of posts before this one I see some very important and respected members of our community questioning or revising their position because of these discussions. I would very much like to ask everyone to take a deep breath and ask yourself why you have been giving our coin so much of your time and effort. I feel confident that once you know why, you can accept how much more easy, even when some other members try to challenge you. To me it seems we still have a long path ahead of us and I'm ok with that. I commit myself to still be here on January first 2017! How about you? For now, a big hug to all and see you on the other side
|
A fool will just look at the finger, even if it points to paradise!
|
|
|
veertje
Legendary
Offline
Activity: 952
Merit: 1000
|
|
December 31, 2014, 12:08:16 AM |
|
I'll continue to follow the thread, but I probably won't push for radical changes like I have in the past. I'm just going to sit back and watch for a while.
I say, keep posting when you feel enthousiasm and there is good news Keeps spirit high. You are great support, so always if you feel you have something to share, don't feel shy because of these mistakes
|
|
|
|
sven22
|
|
December 31, 2014, 12:08:30 AM |
|
I'm getting confused, did the Criptoe team made some thinking errors?
Yes, there were errors in the code I submitted. There are 2 that I can discern from my talks with /GJ today. The first was the 25% limit that wasn't a 25% limit. It was a validation check that acted like a limit on our testnet by rejecting blocks on the confirming nodes. See /GJ's post about this, although all the a/b/c stuff totally lost me. The second part to that line of code was that if it was added to the network now, it would have cause the chain to not sync properly on a resync. Essentially the existing blocks higher than a 25% retarget would have been rejected. So it is my understanding that the results that were shown with the charts were correct because they were valid with a fresh chain on our testnet. However, they wouldn't have been valid on the main chain. Had we worked with the existing blockchain data, we would have identified the error immediately. My team missed it, and I sincerely apologize for it. I take full responsibility for these errors. I knew the code wasn't a final draft. That is why when Rijk asked me to post the git pull, I asked him to have a few pairs of eyes on the code before we did anything. It's also why, in light of the simulator timeline /GJ presented, I asked that we test against the simulator. /GJ is currently working on cleaning up the DIGI code. I have offered to reach out to the DIGI dev in the meantime. However, /GJ is confident that he will be able to provide a streamlined DIGI solution, so I am going to stand behind that decision. I'm going to take a step back from NLG for a little bit. I just need to refocus on what I'm doing. I'm not cashing out, and I'm not shutting down the pool. I will continue to sell physical coins and try to figure out ways to make NLG better outside of the blockchain. I just feel with the animosity over my involvement, and the movement towards finally fixing the algorithm problems, I don't need to be in the forefront for a while. I'll continue to follow the thread, but I probably won't push for radical changes like I have in the past. I'm just going to sit back and watch for a while. Cheers, Fuse Hey Fuse, don't take it to heavily. You're one of the guys who make NLG better. Maybe if not for your/Criptoes work we didn't even had this discussion. Maybe this was required to get people "focussed" again. Myself will keep spreading the word. Some of my friends are allmost that far to join into NLG and now it is still cheap! So if I understand everything correctly, DIGI is off for now? @Fuse, keep up the good job as you're doing one hell of a job!! <3 For the rest, a happy New Year and cya next year
|
NLG Donations: GarNBSDLovSQgPKqhLSiSwGes85VrESgto
|
|
|
Sharkzz1
Sr. Member
Offline
Activity: 880
Merit: 251
Think differently
|
|
December 31, 2014, 12:18:39 AM |
|
I'll continue to follow the thread, but I probably won't push for radical changes like I have in the past. I'm just going to sit back and watch for a while.
I say, keep posting when you feel enthousiasm and there is good news Keeps spirit high. You are great support, so always if you feel you have something to share, don't feel shy because of these mistakes Fuse, everyone makes mistakes, don't forget that, and you are one of the most influential persons here at NLG, small chance that we would be talking about DIGI if you weren't here, trust me. keep up the good work!
|
|
|
|
Halofire
|
|
December 31, 2014, 12:33:59 AM |
|
FUSE!!!!!!
I most definitely appreciate your work, input, and methods! Please don't take too much time.... Maybe get a simulation running while you chill? Nite's DGW for haha's? People respect you here and we all make mistakes so don't get too down on yourself. Take a moment to recollect yourself and come back even better and fresher, buddy!
Best wishes to you and your family for New Years!
|
OC Development - oZwWbQwz6LAkDLa2pHsEH8WSD2Y3LsTgFt SMC Development - SgpYdoVz946nLBF2hF3PYCVQYnuYDeQTGu Friendly reminder: Back up your wallet.dat files!!
|
|
|
ny2cafuse
Legendary
Offline
Activity: 1582
Merit: 1002
HODL for life.
|
|
December 31, 2014, 01:39:44 AM |
|
To clarify, I'm not taking it personally. I just feel with the movement we gained this week to making things work properly, that after 4 months of rallying for it, I can take a step back and refocus. There are a lot of things that I put on the back burner for months to focus on issues here. Now that things are moving, I can step back and see how it turns out. Mind you, if I'm asked to help, I will not deny my help, or my team's. I am, and they are, still fully committed to this coin's success.I just feel it's time to take what I learned here and reflect. If I can apply the knowledge to a future project, it was worth the effort. Plus I've got a new baby that I'll admit comes second to these long posts sometimes. I need to stick to the priorities So happy new year if I don't talk to you all in the next day, and here's to a new year of success! I'm thinking an anniversary coin maybe? Who knows. Cheers, -Fuse
|
Community > Devs
|
|
|
WaterLooDown
Legendary
Offline
Activity: 924
Merit: 1000
|
|
December 31, 2014, 05:54:14 AM |
|
To clarify, I'm not taking it personally. I just feel with the movement we gained this week to making things work properly, that after 4 months of rallying for it, I can take a step back and refocus. There are a lot of things that I put on the back burner for months to focus on issues here. Now that things are moving, I can step back and see how it turns out. Mind you, if I'm asked to help, I will not deny my help, or my team's. I am, and they are, still fully committed to this coin's success.I just feel it's time to take what I learned here and reflect. If I can apply the knowledge to a future project, it was worth the effort. Plus I've got a new baby that I'll admit comes second to these long posts sometimes. I need to stick to the priorities So happy new year if I don't talk to you all in the next day, and here's to a new year of success! I'm thinking an anniversary coin maybe? Who knows. Cheers, -Fuse Hey Fuse, At a few points in time Rijk and Geert asked for help and input regarding these changes. You and your team put your hands up and got to work on a solution, this is what counts the most and was definitely a learning experience for everyone. Once again I personally thank you for joining Guldencoin as you have added so much value. To everyone else have a great new years eve, my 10 month old is sick so I will be staying indoors , so to the rest who are going out to have some fun BE SAFE and chat in the new year!
|
|
|
|
veertje
Legendary
Offline
Activity: 952
Merit: 1000
|
|
December 31, 2014, 09:31:04 AM Last edit: December 31, 2014, 01:00:34 PM by veertje |
|
Very nice indeed to see women can read about Guldencoin there. The women will make crypto work for sales in the future and accepted on mass. Remember the dutch proverb "Vrouwen bloot, handel dood". That means with hot weather, no sales. Women do most of the shopping (it's their hobby ) and with hot weather the women don't buy and go to the beach. So in crypto (with Guldencoin) we need women also to make the sales as well for the merchants. Focus on them as well. When the apps become more userfriendly (read womenfriendly) and initiativs as sendchat and checkcoin will work, the women will come in. Listen to the women in designing apps how they want to use it and crypto will do a big step in the near future. Ang Guldencoin as well. That's why the article in http://www.techgirl.nl is so good to see.
|
|
|
|
bram_vnl
Legendary
Offline
Activity: 1148
Merit: 1000
|
|
December 31, 2014, 09:49:35 AM |
|
happy 2015 to the guldencoin community!
|
|
|
|
bram_vnl
Legendary
Offline
Activity: 1148
Merit: 1000
|
|
December 31, 2014, 09:52:23 AM |
|
|
|
|
|
|