Topic: [ANN][SIM] Simcoin - A Simple Coin - page 14. (Read 157763 times)

@Pentamon

Sim-Ipo is over, Ora "IPO" is over. Let me here your schedule to bring them in one boat. But a lovely dream indeed

Look in how many troubles a simple blunder of allowing any account put NXT: https://nxtforum.org/nrs-releases/nrs-v1-2-4/40/

That's why you need to think, think, think and don't be like "I'll just be lazy and make some crap fast and then we'll figure it out".


My design goal is "at least 100 transactions per second with confirmation time under 1 second".

By the way, transaction size:
* Bitcoin: 250 bytes (although it can easily reach 5-10 Kb, depending on the number of inputs and outputs)
* NXT: 160 bytes
* SIM: 96 bytes

Not bad.

Sometimes, I daydream:

I think:
On the one side, there is Simcoin, a wonderful, dedicated, serious software developer, doing his thing alone. Methodical.
On the other side, there is Ora: a wonderful community of people, with no software. Developing a project plan. Methodical.
Could they get together?

That would be amazingly powerful...

And then I wake up.
Pentamon

What speed of transactions it will have?

Here's a broad roadmap for Simcoin:

I would like to finish the design and implementation of the core in August/September. Not sure if that's enough time, but that's what I'm aiming for.

When it's done you won't see anything big, because all the changes will be under the hood and the test client will look pretty much the same. But I might release some technical details of implementation for discussion and review, so you can peek under the hood if you understand crypto tech. Also parts of the source code will probably be released for developers and people hunting for security bounties.

To finish the core I need to:
- finalize the improved consensus and related code refactoring
- finish bootstrapping (it's mostly designed, just need to code it), so nodes could go offline and then catch up
- finish client protocol (this can take some time, depending on client's feature list. Also might need to implement these features in the test client to be able to test them)
- implement efficient subscription for clients that can handle millions of accounts

Depending on how much time is left, I would also like to implement in nodes:
- the reward tool
- voting system (it's an essential feature in decentralized systems, because they need a way to self-govern)
- flexible account control (for improved security and multisig)

Then we will need a stress test tool to make sure the system can scale and work reliably under a very heavy load.

This is where the test net will be launched and I can start hiring people to work on the real client and all the other tools and projects. At that point things should start moving much faster and hopefully you will see something "big" around Christmas. Meanwhile I will continue improving the core and adding more features and tests to it.

The official release will probably be somewhere in the spring.

So this is a rough plan, assuming nothing unexpected happens

Thanks for the update.
Will we see sth big at Sep?

Progress report


(guess, no ice cream for me...)

Another slow week, although things are starting to pick up:

- Continued developing improved consensus scheme.
- Network packet rounding replaced by optional junk at the end to decrease node traffic and allow better protection for clients.
- Bug fixed: one field from packet header was not encrypted.
- Account registration request is now a transaction.
- Switching to the new consensus scheme allowed to refactor transaction handling code. This, coupled with changing new account registration request to be a transaction, helped unify the code and it now uses a single transaction handling task instead of individual tasks for each transaction type.
- Trying to unify network requests for transactions as much as possible.
- Bug fixed: truncating second layer of encryption part on new account requests now requires a different way to detect it.
- Finalizing design of bootstrapping/account info requests.
- Updating wiki documentation.
- Accounts class to cache ledger accounts in memory. Simple memory pool for more efficient allocation/freeing.
- Client: adding vault accounts into the memory cache on each JS API load_vault() call.
- int5/int6 classes to handle 5- and 6-byte integer fields transparently.
- Thinking about account control.

No new version, because it looks exactly the same, just works a bit faster due to improved consensus. So no reason to waste time updating the nodes.

Nights are getting colder, woo-hoo!

The main idea of Ripple/Stellar is currency bazaar, so I'd say Simcoin is closer to Bitcoin, because it doesn't attempt to deal with multiple currencies.

@Nxtchg - In your opinion, will Simcoin compete mainly with Bitcoin or Stellar/Ripple in the future (which market is more suitable for Simcoin)?

Thank you for the status report, keep up the great work

Progress report

Not much progress this week either



These are probably the hottest weeks of the summer. I expect temperature to start dropping in August and my productivity raising back again.

I am still working on the improved consensus, it's about 90% done and works on my local machine.

No new version though, because I want to implement the subscription model for clients first, otherwise polling interval kills any confirmation speed.

See you in a week.

Updates are posted every Sunday.

Why are people so impatient?...

Thanks, and welcome back. 

Progress report

No progress, took the week off.

I was away for 3 days and unfortunately web server on the exchange froze just after I left.

It's back online now.

In my official capacity as puny humans communicator I have no information to relay. I just hope he is okay. For anyone worried he abandoned the project I say there is less than .00004% chance he did. If he did I'll eat my hat. Imo he is probably acting like a mad scientist, unshaven and forgot what day it is. He'll log in after he remembers he hasn't eaten in 2 days.

Good for you. Meself, I lowballed too low...

Simcoin feature request: multisig.

Bitcoin has 2 of 3 multisig.  I would like to see Simcoin have a more powerful version of this: N of M multisignatures where both are specified at the time of address creation with M keys and a N in the set {1 .. M}.

Thanks for considering.

The prices of the last days allowed me to also become a member of the sim community.

Looking forward to test the client  

Distribution update



255 accounts own simcoins now vs 210 after the IPO.

The top 10 accounts own 21.25% vs 21.00% two weeks before.

Progress report

This week my primary HDD died. It was 5 years old and every summer worked in excessive heat. Then the HDD cooler stopped one day and that was the last straw for it.

Fortunately, I had a clean XP installed on SSD and the drive didn't die completely (thanks, Seagate!), so by cooling it down I was able to copy everything to a new drive. Then XP booted as if nothing happened! Whew...

Still took two days out of this week

Did:
- Fixed a bug with account info task that didn't reset the response queue properly and so caused slow balance updates in the test client.
- Added 5th test node.
- Major refactoring of consensus code.
- Client: added Ctrl-C/Ctrl-X/Ctrl-V shortcuts.
- Client: fixed account auto-unselecting, due to updating by timer.
- Minor fixes, improvements and testing – seems stable and reliable now, even with small network and a lot of checks and code missing.
- Tried modified time sync version: the time drift is much slower, but still continues. Have another idea, will test it as soon as enough time passes – I want to see where the current timestamp will drift. Currently it's about 15 seconds off the correct time. And agreement between servers is 1 second. Not bad at all considering servers have 1 minute time difference.
- Setup a new test/development server at Forked. Both servers, by the way, had 100% uptime so far.
- Two new ideas how to significantly improve consensus and at the same time decrease network traffic. Can't pass this opportunity, but need to write more code unfortunately. Still, since consensus is the core of the system, I believe it must be done.
So this is what I am working on right now.

Sorry, but the slump in SIM merely mirrors the slump in a lot of gen-2 cryptos. The man you love to hate had nothing to do with the fact that an excited investor in Qora at 75 satoshis or an excited investor in NTX at 30 satoshis is riding butthurt after seeing his stake drop well over 50%.