Thanks a lot for the explanation!
What can this be used for? Or what’s the benefit?
2 Likes
Not being greedy here I promise - is there any more clarity on the possibility of DBCs at some stage now that things are getting clearer? What might it look like, if so?
Phenomenal work as usual, it’s a pleasure and a privilege watching this complex beauty come together.
7 Likes
Privacy for one (bad elders won’t be able to progressively canvas transfer histories to build a record of clients’ transactions). Another one is that it would keep the size of the transaction history small by essentially pruning it.
4 Likes
gogogo team! its REALLY getting close!
4 Likes
Thanks for this summary. Good to have all this in one place!
I find the need for lazy messaging bit unclear though:
I am a bit lost with all that. Could you tell where the lazy messaging is going to be guaranteed i.e. where it is needed before there is any point in releasing the testnet?
Also, is there anything we as a future participants could do in preparation?
1 Like
There are 40 coming weeks in 2021. It could be any one of them.
3 Likes
LM Solves the “in sync” problem, so as the network infrastructure is changing, nodes leave/join etc. LM keeps everyone securely up to date in 100% of messages and events. It is really a catch-all error and fix in relation to infrastructure. To get a testnet out quickly we might not have LM in all of our messages and accept some level of errors. It’s a real tension for us in-house, get the testnet out quick or get it out with a hopefully small amount of error.
We feel it’s been too long and we will go for it for now, so testnet out, see if there are any other issues while we complete LM throughout. It’s not only implementing LM it’s also making all messages more uniform, there is a difference right now in routing messages and node messages and we want them all to act the same way and be clearly understood.
tl;dr Many networks don’t have LM, like IOTA when they lost their network LM would likely have solved it, but it took them a while to get the irrecoverable error. LM is simple and very effective, possibly new to the world, but we can still test with it partially done, to check other areas of the system.
The other wee tension is data types, they are close to completion now, but again we figure, push and test.
Even with all that folk will still say “perfection is the enemy of production”, “I hope you are not over Engineering” and many more “smarter than you” statements 
I am ready to read them .
16 Likes
OK, this clears it up. I was just wondering, because the message was confusing.
All the more power to you 
8 Likes
Please try to work it in. Immutable only is just fine.
3 Likes
Is the web browser going to be working for this testnet or is it CLI only?
Very much an outside chance, but CLI for sure.
3 Likes
Even if its not its good to hear it wouldn’t be too long down the road of updates…which is awesome.
6 Likes
Who said 2021? 
This is a very important aspect.
If data fails to scale up, lots of useful services will be impossible to implement.
I hope that such sort of concurrency at least possible.
There will be no way to make a limit on operation count, like it was done in previous releases to hide this problem.
3 Likes
Bounded Counter in POC as we speak, it’s all good.
9 Likes
You still need to have a public ip and not being stuck behind a isp carrier-grade NAT?
We hope not @tobbetj. I have been able to connect to testnets from behind an ISP NAT, i.e. no public IP, as have a couple of others in-house here who were not able to do so previously. Like I say above though, this is a tiny sample size so we’ll need to wait to find out once this is live.
Something to note is that some people who were behind a NAT were able to connect to previous testnets. It wasn’t quite clear cut that you could only connect to them if you had a public IP.
6 Likes
That sounds very good and promising. As you also touched on, there are probably many people behind a CGN, so if it works that would help many users.
3 Likes
I wonder if there will be an efficient way that error messages can be collated this time around.
How many error messages exist? I’m guessing eventually, the self healing aspect kicks in based of these errors being self reporting.
Is it possible/ practical to build a database of these errors and expose them via web address ala Google search style. Copy/ Paste an error message into a search box, it’s matched and you guys get a dashboard of hotspots.
Maybe there will be zero errors 
5 Likes
Yes many do they are errors saying I need more info and you have that so resend your message with that extra info. Some of course are actual errors, like that data you asked for is not on the network etc.
Good idea, yes we have all the errors defined in each library, so we can extract these.
7 Likes