Am not sure exactly how this would work but I do like the concept of redundancy.

So I predict that in the future the vast majority of stake holders will not have much technical expertise and will simply have given their forging rights to a pool and we are likely to end up with a small number of large pools (say 10).

If those pools are attacked in any serious manner then the 1000+ TPS network will be slowed to a "crawl" with a lot of non-technical stake holders probably blissfully unaware (too busy enjoying cocktails on a tropical island).

Even if they become aware reasonably quickly they won't have enough computing power to "bring the network back up to speed".

Doesn't any else see this as being a problem?

How many major pools are there for Bitcoin?

Understand that those pools have hundreds (and maybe even thousands) of "hashers" who can simply, quickly and easily "change pools" if there is a problem with on or even start mining for themselves if the pools were being "shut down".

So the Bitcoin network can quickly repair itself even if *all* the pools were stopped.

I am worried that Nxt will not be able to do that if a majority of its stakeholders aren't even concerned with running the software.

So shutting down the pools in Nxt could be a much bigger problem IMO.

As far as I am aware you do not have to be online for pooled forging but perhaps CfB could confirm/refute that.

Granted it does do that (so yes - saying *zero* benefit was going too far).

My main concern is that simple economics will likely lead to a small number of "super nodes" because I think *most* stakeholders will lease their forging rights to pools (why wouldn't you want to get money for doing absolutely nothing?).

So the problem is that if those pools are suddenly shut down (due to court order lets say) then the transaction confirmation rate is going to slow down *dramatically*.

It would look rather bad if our amazing 1000+ TPS network suddenly ended up struggling to handle much more than 10 TPS because it was actually now only being run on a few hundred Raspberry Pis!

Not only is there no reward but under the currently proposed TF there is actually *zero* benefit (i.e. they are not helping to secure the network at all).

Maybe I need to draw a picture (as clearly people are just not getting this).

We are going to have this situation after TF is implemented as currently planned:


Where 1/x is each pool's chance of forging the next block (once we have pools then non-pool nodes are going to be a very small minority).

You are only going to send you tx to one of those pools - you are not going to bother sending it anywhere else as you want your "instant confirmation" (and who wouldn't?).

So the network is being *secured* (meaning forging blocks not just passing packets) by a total of "x" nodes.

Clear enough yet?

Of course people can run a node with zero chance of getting any forging reward but they are not helping with confirmations (so they are not really helping to *secure* the network at all).

Understand with TF we are actually going to be directly sending our requests to probably 1 of x servers (where I expect x will not be a large number) owned by pools (so the rest of the network is actually not doing anything apart from trying to "keep up").

In regards to forging we actually have a dilemma - so let's look at it carefully:

Current approach of PoS with "pools" will lead to exactly the same situation as Bitcoin (a few pools controlling the entire network). This is okay from a performance perspective (you'd expect them to be running great hardware) but it is not okay from the perspective of decentralisation. If governments decide to seize the servers of the pools then Nxt could literally disappear.

The approach I have suggested will probably not be "botnet" proof - but at the end of the day does that matter?

At least it is decentralised - and if everyone is arguing that the whole "forging" thing is something we shouldn't even talk about (as it is not an incentive) then why not let the botnets help us to secure the network?

To use a Simpsons style quote:

I for one welcome our new botnet overlords!

It is just software so *changing the rules* will only be an upgrade (optional of course).

The idea is simply to make it more likely that even a forging account with 1 NXT has a pretty good chance.

We can *limit* the accounts and make them only available to previous stake holders if required.

(i.e. there are solutions to this problem)

My idea does not depend upon your amount of NXT.

I doubt it also - but I do think we can address the Sybil attack vector much as with "hallmarking".

I am only trying to *help* but for sure my idea might not be the best one.

True - but it won't help if you have to "register" those addresses.

The purpose of my proposal is to "change the odds" (in favour of the small stake holder).

No approach we take will be "perfect" but I think if we can make it "hard enough" then it will result in a "fairer system" (that is all that is being proposed here - and it is just a "proposal" so unless enough of the community agree with it then I doubt it will be implemented).

EDIT: For the record I am *not* one of the 73 (but also am not *jealous* of them - I actually thought Nxt was a *scam* when I first saw the post).

Another update - I've removed four op codes (ADD_VAL, SUB_VAL, MUL_VAL and DIV_VAL) so that the only op code that requires a 64 bit value is SET_VAL.

Other changes were in the description about the stacks (there are two - a "call stack" and a "user stack").

Yup - this seems like a good start.

Maybe so - but I think we could make it harder for them by requiring accounts to "register" their IP address.

(so likely they would end up stomping over each other)

EDIT: If we limited "changing" IP addresses to once per day I think that would make it even harder for botnets.