The amount of memory usage is linear depending on the number of blocks in the chain and the number of transactions in your wallet. I actually have a wallet with over 36,000 transactions and it has a total of 7 loaded wallets running and staking 24/7 and running S4C. At the time of writing this it is using 745,000K. Another one with only 259 transactions is using 359,000k
All wallets operate in the same way. Bitcoin has changed this slightly by doing some things regarding when to load a block into memory. It will load them as needed. But does not unload them. So it grows and grows until it is restarted. HBN has 30 second block time so it is getting more of a hog as time goes on. I will be doing my best to address this going forward, but there will only be so much I can do.
This is also only a slight concern of mine, as the memory usage of HBN should grow linear as time goes on, but memory size and speed will continue to grow exponentially, as well as the cost decreasing exponentially. So eventually it will not matter much.
But can be a turn off for some people who have lower end computers, or run many wallets.
I'm wondering whether HBN's emphasis on encouraging small stake values is contributing to the problems spoken of here. The maximum coin block size to optimise Proof Of Stake is 250 HBN. HoboNickels consistently refers to this reward as a '~2%' stake once a coin block matures and not '~2% interest'. The reason given for this policy is that it will keep people in Stake mode longer - 24/7 for some - and this will help with network security.
I suspect that this might not be as effective as might be thought. The average number of active connections on the HBN network isn't that high, and some POS coins without maximum, coin-block, optimum stake sizes perform quite well by comparison in that area. I suspect that the difficulty in staking turns quite a lot of people away, who would otherwise significantly add to the security of the HBN network.
To have stake-splitting combined with predominantly small coin block size values leads to many, many, smaller calculations. Like a chain reaction. I should imagine that is a major contributor to the blockchain expanding. Is not blockchain saturation a security risk and a vulnerability for a form of attack?
Recently, I completed a full installation of the new HBN wallet on a 32-bit Windows XP machine with 8 Gb of RAM. It took nearly three days. I suppose that is to be expected for a wallet with a long-established blockchain, and HoboNickels has been around for longer than most. I have a faster i5 64-bit machine. The HBN Windows client is tagged as 32-bit. Will it run on a 64-bit machine with a 64-bit Windows7 operating system just as easily?
I found that the old, about to be phased out, wallet (v4.0.1) performed well. Transactions were fast and it was quick to load, although rather slow to stake. This has been my experience since first coming to HoboNickels in April. I hope that the fast stake v5 will be better. I've split 26,000 HBN into 10 coin-blocks of 500 HBN and the rest in 250 HBN coin-blocks - all with individual addresses to help keep track of staking. The 500 HBN and 250 HBN values were based on previously given advice.
I suppose what I am really trying to get at here is that I am sold on HoboNickels as a concept, but have some reservations. I would welcome stake-splitting with far higher parameters, if not phased out entirely, and the raising of the maximum optimum stake level to a much higher value, if not abolished altogether. I do not want to have to run my computers for long, extended periods in order to generate Proof of Stake. Certainly not 24/7. It doesn't fit in with my lifestyle. I prefer POS to be more like interest - there and easily accessible as of right - as opposed to being like the laborious, Proof of Work with a questionable break-even conciliation. That is too much like mining. I have better things to do!