dont quite follow you on this one




lol new guy missed out on all our previous 'fun'

I just want to make sure we can implement the fancy cross chain algos, dont care about details, just need to be able to do atomic cross blockchain transactions

For 1 NXT = 1 vote, you need to:
- only accept vote from NXT on an account just before start of a vote (alternatively, just before first vote). This moment in time is now fixed, let's call it t.
- after decoding a token with a vote, check the balance for account at t.

That's it.  

Adaptive bandwidth usage will be very important
We can have infrastructure that can handle bursts of very high tps, but only require lower sustained tps for all nodes

This should get R least two to one Advantage, maybe more
If moores law doesnt take us all the way to desired tps, we can always do the parallel blockchain approach
So both methods and we can be adaptive, ride moores law and have parallel blockchains

Details,details
Where is infrastructure committee?

Don't know which posts you mean. But I followed the discussion for a long time now.

And still not convinced that 1 user = 1 vote is best way to vote if that very user has nothing at stake during voting. So, just voting because he can and has actual no real interest or knowledge in the topic.

Why? Are we short on bytes in nxt client?

Again, why? That's just a completely arbitrary restriction.

Why not?

Even Intel (and AMD) disagree with you, what with AES implemented in the cpu... and this is supposed to be an interpreted language where everything is possible and permitted

I maybe repeating myself; read my previous posts.

It's not that would not be possible, theoretically, on the paper.
But, somebody has to implement it. The less, the better. The less, the less bugs, the less security issues, the less maintenance effort etc.
We discussed that issues at lengths some days ago.

Let high-level libraries handle the complex stuff.


Maybe, but it's a reasonable one. It simplifies a lot. See above.

Not because it can be done, should mean, it should be done.


A low-level. So, each thing, you can build upon the very basis should be done on top of it. Not besides it: layered architecture.

It's just a pattern in computer science. It works quite well. Why violating it? Just because we can?

Seems like it, but we might end up with forks on basis of client version.
I've just seen someone with a heavy hallmark running on 4.7e......god knows how that still works.

@ChuckOne and Panda:
One of the good things about having the voting site is that we can play around with various voting parameters until we come up with a system (or systems) that upset the least number of people for the shortest time.

It's never going to be perfect, because everyone is slightly different, and yet everyone thinks that their own viewpoint is the right one.....democracy, eh, what ya gonna do?

On the forking front, I get the feeling that 5.12 likes to have a fresh blockchain. I've updated 3 NRS to 5.12, the one that downloaded a new blockchain is running fine and on the main fork, the other 2 (with blockchains from 5.11) forked off in very different directions.

Did the author really attribute NXT's success to the ~60 members of reddit? Terrible article, it reeks of paid advertising.


Now that you mention this, I do remember C-f-b mentioning 32GB would be the hard limit, but why is this?