Topic: [ANN] SMLY- smileycoin - for rewards in education - page 25. (Read 77288 times)

Maybe the instructor should get the uncollected coins.

Now there's an incentive program that could help SMLY!

Usually they do the exercises because an instructor tells them to  

So currently most students aren't taking possession of their SMLY? They're just doing the exercises for the educational benefit and/or fun?

The functionality of rewarding with progressively more SMLY as the student finishes more lectures can be done purely within the tutor-web system, i.e. without knowing the payment SMLY address. Until now we have just had the rewards inside the system until the students choose to redeem their SMLY.

But we've learned that most students just ignore the SMLY (which is why we've set a deadline, after which the SMLY just get donated).

Your idea to link the rewards/amounts explicitly to a registered SMLY address gives an added benefit: It would allow us to reward with some bonus those students who have downloaded a SMLY wallet. So we can direct increased SMLY rewards at those who have shown an interest in SMLY.

Cool, really cool.

Unless someone sees a problem, we should just implement this!

That's a new one: One could give minor rewards for the first few lectures and then increase them as students complete more material.

This is definitely food for thought!

Perhaps another way to control fake accounts would be a sort of PoS+PoW (where the 'W' in PoW is Work performed completing exercises). It could be something like this:

-student logs in by demonstrating ownership of a particular SMLY coin address
-the reward they earn for completing a task is proportional to the amount of SMLY in the address

The problem with this is that it would also incentivize hoarding, meaning nobody would spend their SMLY at the University cafeteria - bad!

So, instead of basing the reward on the amount of SMLY in an address, the reward is based on the amount of SMLY the address has previously earned from Tutorweb irregardless of whether that amount is still in that address.

No, haven't thought of this - yet 

But a kennitala or social security number wouldn't really work since we have a number of students e.g. from Kenya.

Of course we would love to see students from various slums deciding to use the system, but as far as we know, we have only had low-income students come in through various teachers in places like Kenya. Some of these instructors have just been in touch and we've set up a class for them.

There are a number of student registrations but we don't have a good overview of how far they get.

Have you thought of a way of controlling registrations through the physical distribution of anonymous accounts to students, or perhaps using kennitala as in the Auroracoin airdrop?

We've seen this a few times (we manually check the rewards on a daily basis) - and we've added sentences describing that this is an abuse of the system. The first thing is to send them an email, which is easy enough. If they continue we've reduced the reward, which is awkward but not really a problem:

The principle should be to have low rewards for a unit with small and easy content. A "unit" here is a "lecture" and when people ace it they get the maximum reward for the lecture. Lectures are grouped into tutorials and when someone aces an entire tutorial they get much more.

We have a goal of spending a certain amount per year and to get there we had to increase the lecture rewards across the board. The fact that the abuse cases lead to a reduction in a handful of lectures is therefore not a problem - it just means that some of the easy lectures have much lower rewards than the more difficult ones -- and much, much less than the tutorial rewards.

There are a few more changes on the agenda: For example, students are sometimes asked to submit new material, as new drills or as examples. We will implement high rewards when students submit good material, as determined through peer-evaluation. This is much harder to abuse and is worth much more since it also improves the system. When we get to this stage we can reduce the generic lecture awards across the board and pay more for the stuff which actually shows you are really good.

Of course we want to keep the system open. Although a user can not do the same drill multiple times for rewards, a person can register under different user names and this is rather hard to limit without reducing free access.

How do you deter students from gaming the system by completing the same easy quiz over-and-over for rewards?

So March is almost over and there have been no objections. Also we have secured funds and people to implement stuff.

We will therefore try to implement the following changes in summer:

• Add X11 as a second mining algorithm
• Change the 10,000 SMLY block reward allocation: 1,000 to the miner; 4,500 to large SMLY holders; 4,500 as donations

A large SMLY holder is any address holding at least 25 M SMLY (i.e. untouched for some time, initially about 1.5 hours).

A collection of (10?) rotating donation addresses will be in the hands of the non-profit organisation Education in a suitcase (normally as cold storage). Suggestions on how to use the donations can be raised on this list.

It sounds great! 

OK :-)

Re "oldest block": Suppose miners are mining block 131 and we both have 25 M SMLY, you in address A and me in address B. Suppose also the content at your address A was last changed at block 120 but mine at block 125. If everyone else has changed the content of such addresses after those blocks, then you should get the next "interest" part of the block reward and your "most recent block" changes to 131. When block 132 comes up I will get the reward and my "most recent block" changes to 132. Someone else will now have the oldest "most recent block" and get the interest from block 133. Eventually everyone has had their "most recent block" changed and yours becomes the oldest.

If this is doable, then it should give a fair way of splitting the reward in a PoS-sort of way. The algorithm just loops through every "holding" address and rewards them in turn.

Anyone who changes the amount in an address just gets moved to the end and that's no big deal if it just happens occasionally, but if the amount in the address is bouncing all over then this is not a "holding" address.

I should add that this is not a new idea. Dash does pretty much the same thing according to this report: https://cointelegraph.com/news/fastest-crypto-for-cash-ever-dash-funds-wall-of-coins-integration

-- and that should imply that this is doable, one way or another :-)

In the next month or two we should define exactly how we want to modify the mining algorithm and, if needed, the reward scheme.

It's really tempting to move towards a multi-algo coin. Since we only have finite resources it's also tempting to just pick a well-established second algorithm, where X11 must be a strong contender. One worry is the existence of some X11 ASICs. It would be good to get a discussion on this.

Next, we should think about whether such X11-generated blocks should have exactly the same reward type as is currently used. An argument could be made that we should do one or more of the following:

• Auto-distribute some of the block rewards to existing coin holders.
• Auto-donate some of the coins to a specified cause.

At one extreme, the miner gets very little and there will be few miners. At the other extreme we're just doing the same as before, with one more algorithm. In-between, we may be encouraging people to hold onto their coins and we may also be supporting something useful.

Let me give some background. If I were starting the SMLY today I think I would change at least the following:

• Choose a different mining algorithm (or a combination of several).
• Try to give an incentive to hold onto one's coins. I'm not convinced that the proof-of-stake approach is the way to go. My preference would be to figure out a way to take a large portion (30%?) of the mining reward and simply split this among existing owners.
• There would not be a premine.
• One would automatically funnel a large portion (30%?) of the mining rewards to a handful of addresses belonging to specified charities (the tutor-web system and the Education-in-a-suitcase, EIAS, non-profit).

In the current case there is no need to donate to the tutor-web system which owns the premine (and we can't really take that back). It would be extremely useful for EIAS to receive donations.

Similarly, suppose we figured out a way to maintain a list of all addresses containing at least x SMLY and we split 30% of the mining rewards evenly among those addresses. If this is a part of the criterion for accepting a new (X11) block, then my gut feeling says we won't have any ASICs messing up the scene. It also means we have a strong incentive to keep a whole lot of SMLY (we can discuss the numbers).

Another way to say this is that the new (X11) miners get a lower reward than the scrypt miners. Of course that will result in fewer miners bothering to mine, but it will also result in a lower difficulty.

I don't know whether this is all feasible or acceptable, but I'd like a discussion, please!

The cryptopia mining pool seems to be closing down. This was a very, very nice pool: If you have any influence at cryptopia, please help us to keep this running.

if it will be so ..then it will be a good chance to flourish a project on right path .....