Especially if a witness is not updating its price feed for days, then there is NO reason for that witness to be voted for.
Not true.
Why would you say something like this?
You are viewing a single comment's thread from:
Especially if a witness is not updating its price feed for days, then there is NO reason for that witness to be voted for.
Not true.
Why would you say something like this?
Well, if you are stale for months and without price feed updating for months/years... what's the point?
But let me know your view...
I am signing the blocks alright, but I don't update my price feed.
https://peakd.com/witness-category/@felixxx/price-feed
...maybe... just because things work... it's not why all is 100% alright?
From your comment... on your post.
Why you say this? Where in the code you can find this?
I can actually show you where and how the average is calculated.
Do you want me to? Will you edit the post and stop spreading misinformation?
Yes if I agree with your answer...
It says so here:
https://gitlab.syncad.com/hive/hive/-/blob/master/doc/witness_parameters.md#L9
However, here's the median price feed:
https://gitlab.syncad.com/hive/hive/-/blob/master/libraries/chain/database.cpp#L4310
for( int i = 0; i < wso.num_scheduled_witnesses; i++ ) {At first glance, that actually looks like it could be 21, not 20...
I looked into where that
num_scheduled_witnessescomes from, but that also looks like it should be 20:https://gitlab.syncad.com/hive/hive/-/blob/master/libraries/chain/include/hive/chain/witness_objects.hpp#L171
Anyways,
I have pointed you to the code.
...it always said in the official repos that only the top 20's price feed gets used.
I can't see where that happens, though. I trust it still does and I am not reading the function right.
Ok, first thank you so much for going look into the code. That means a lot to me.
Let me try to clear the confusion here. And I will amend my comment so it becomes more clear why it matters (I wanted to be a bit harsh but I guess I need to be more correct and less influential) the higher you are as a backup witness. But you can do your own research and find out more if you wish.
HIVE has so-called "rounds" of scheduled witnesses (20 top + 1 backup). During this round, the scheduled witnesses will produce blocks, one by one, 3 seconds apart. Then, when the round is over a new round is elected/shuffled. In this next round a new backup witness is selected (there is a ranked weighted rotation schedule that can change on every round) and then during that round the new 21 witnesses (20 top + 1 next_backup) another 21 blocks will be produced, and so on.
The price feed median is calculated on every block produced from all the scheduled witnesses in that round. Meaning that every time there is a backup witness on that 21 witnesses round, the price feed advertised by that backup witness will be used for the calculation. Hence, backup witnesses should have a correct price feed.
I agree on the part that it's not the end of the world and this is why we have 20 top witnesses to have a higher weight on this median price. But I still like to think it matters because I am, after all, participating on a schedule eventually. So, that was the reason for me to be so harsh on my comment that I will now amend...
Here is a very cool tool done by @arcange where you can check this in movement... He shared it on this post: https://peakd.com/witness-category/@arcange/witness-schedule-testnet
As you can see you already produced a block and your feed was used when your witness was in the scheduled witness round where your witness produced the block.
As always, discussing and asking stuff makes us learn a lot every day. I am happy to fix/change things if I find I am wrong... no harm in that. And people should be aware that they either do their own research or they can be lured on mistakes/opinions that might not necessarily reflect their own.
Cheers (and don't be so harsh on yourself... that misinformation part was a bit too hard but ok. I was just putting up my opinion... of what I think matters for witnesses under the big top... I don't even want to go that higher because I don't feel I am giving out enough to be that high... and also because my node is mostly for my own usage as you can check in this post).
Done, post updated. By the way, I have added an example of how easy is to setup a price feed... but there are a couple of examples you can already find via ecency search bar for example:
https://peakd.com/witness/@rishi556/how-to-setup-hive-witness-pricefeed-using-hivefeed-js
Ok... as I don't either believe "days" will make a difference in an average... I rectified the comment to "weeks". Which can "most likely" make a huge difference, depending on how many do the same.
Post updated. Tks for all the rollercoster. At least it made me learn more. Hopefully I provoked the same and it wasnβt a waste of time (I would hate that).
The median feed prices being actually protected by the top 20 witnesses if all (or at least most) of them have correct feeds, its a good thing (and u where right about it, I was wrong). I think half way of our discussion I was thinking on average and not medians, hence all the discussion. I blame time.. was in a rush with other personal stuff.
Either way, its a very easy thing to automate and that is why I also never understood why witnesses choose not to do it. Sure its another thing on the ledger⦠but its so tiny⦠and eventually something that can go into a separate chain if needed, freeing the main chain from things like this. Could even rely on other signals, such Dexes or the internal market itself.
Will check you out more in time. I have some stuff you posted in the past that I want to read and research.
Do you have an open RPC (and/or history)? And if yes, how close is to NZ for me to test if I could have better usage of it?
In the post you are still writing that a stale pricefeed was bad for the blockchain.
I am a backup witness and get included into a round every 6 hours.
The price is the median of medians, actually.
If my value is off, it won't be used for any claculations.
I disagree. I am just focusing on the stale ones... price feed up to anyone. But I need to describe the fact that those witnesses are not wishing to be "witnessing" for that long... the best way I can find that is to show for how long they are not updating the price feed.
I can say any non top 20 witness can set the price feed to anything that does not matter but that would be completely contra-sense for the reason that variable exists and also confusing. And I would need to explain a lot more that people will not care to read or would necessarily understand. So, I opted to use the time you are not updating the price feed with the stale witnesses. I think its fare... if you are stale for a long time that matches the time you don't update your witness price feed... I don't think that witness should be voted, period. my view... my opinion. You don't need to like it.
PS: I can add this comment there if you feel that's useful for others that read.
I think people should not vote for witnesses who have a hard time keeping it simple and truthfull. Just my opinion.
π the truthfulness is something that takes time to learn (unfortunately we are not gods)... but eventually (or hopefully) we get there, even if experience takes time to master. The hard time I am having to know you better, so I can learn more, does not even compare with the feelings you create when writing (being in this post or others).
I have nothing against you mate... but you could be a bit less negative and a bit more helpful. My opinion. Also, aggressiveness "might" be your thing (can't really tell yet)... but it's not mine (its my interpretation of your posts, but I might be fooled by not knowing you better, so take that into consideration).
I need to carefully choose my words before addressing you. Otherwise it will not be a very pleasant conversation at least for me. I learned at least. That's what matters to me. But I wished I could just have learned in a more healthy/fun way. Not your kind of business?, maybe... but hey, for me it has value.
I need to rethink with more time about this interaction and learning experience. Talk again in a few days.