Best Selling Option

[monsterer]

@bytemaster - this whole discussion is based on the misconception that 101 delegates would not operate at a loss, when clearly this is not the case and the inception of the bitshares 1.0 chain (and almost all other POS chains) proves this.

You and @stan have already spent the best part of 1 year arguing that 101 is the optimal amount of decentralisation, and now you propose to throw all that work away?

[liondani]

What about use 17 power-witnesses plus 101 regular-witnesses(or better a dynamic number)....
17 that get paid 80% of the total budget and get elected and the other 101(or what ever) get not elected but the code pick them up automatically depended on reliability, active feeds, latency,luck(?) etc. and they get 20% from the total budget...

I think marketing wise its inferior. It seems more decentralized than bts1.0 (but obviously it isn't)


PS1 Of course the powerwitnesses sign 80% of total blocks

[santaclause102]

The only relevant question imo is whether the 'best-selling' option, significantly adds to costs or security concerns.

BitShares is attempting to sell itself in the market and the "Customer is Always Right" so all technical concerns aside we have to market it well.

I think I have made a entirely defendable case that we do not need a large number of "witnesses" for the network to be secure... just like large speakers are not required to produce good sound.   For a very long time people would buy the larger speakers even if they were technically inferior and more expensive.

So rationality is not the most important thing when figuring out how to build something that will sell.

This means we have two options:

1. Figure out how to sell the best technical approach
2. Adopt an inferior technical approach as a marketing gimmick. 

So in the spirit of marketing gimmicks we need to identify the fallacy that others BELIEVE is good, and get them to accept BTS as equally good.

People LOVE the idea of having 10000 potential block producers with no barrier to entry.  But what this really means is that people love having the ILLUSION of direct control.

But perhaps a bigger issue we face is WHO are we selling to.   Altcoin fanatics or the general public.   

If the requirement is to build a FUNCTIONAL system then we have that and have already optimized all of the security conserns. 
If the requirement is FORM over FUNCTION for sales reasons then we should debate what FORM sells the best independent of any security concern.   

Perhaps all we need to do is add a nice FORM over our best FUNCTION in order to sell the best.

Those with the most money are not the ones screaming on reddit and bitcointalk.

Often what sells good is just what is different and also provides great potential for controversy which is very good! Controversy though requires a good command over a subtle yet dumbed down marketing language which we sucked at up to now.

[wuyanren]

Why don't we have a vote, let everyone choose 17 or 33？ :'(Voting deadline should be 7 days in October

[liondani]

The easiest thing to sell is simply a product that just works!  A product with a simple, easy to use interface and rock solid performance! This alone will sell the product and create buzz. *Now I think of it, would it be a good idea to have a "customer testimonial' page with 5 star user rating?*

For those of you that really believe that there is a big mythical market of people that don't care how it works but that it works. Why don't we just run BitShares 3.0 from one computer in a safe haven jurisdiction with a great website and a couple of customer testimonials?

Chances are we won't be shut down for years and can surely become pretty big because the main market doesn't care how it works right?   

----

If you look at Bitcoin demographics

Yes there are some people that don't care how it works but that it works but I believe most users have some concept that Bitcoin is decentralised by thousands of computers worldwide mining for Bitcoin & processing transactions.

Then there is a smaller group of people who know Bitcoin centralizes into pools but believe hashers will switch if needed/incentives are there/6 years proven model. While  there are only very few who understand the Bitcoin security model is very weak.

With 17 witnesses you would be creating the reverse, only the small group at the very centre would believe it's secure & adequately decentralised. We know some in this community and definitely the alt-coin market would believe it's not sufficiently decentralised for a variety of reasons. (While I believe the referral programme and partnerships are strong but I don't think BTS is going to expand so rapidly in the next 6 months that you can ignore the alt-coin market.)


I also believe while the mainstream market can be convinced to put their money in a widely decentralised system, they would view one processed by 17 people as too centralised too.

 "Instead of using a centralized trading exchange, that offers great customer service, is insured by Lloyds of London and fully regulated, you can instead use BitShares which has none of those benefits and is run by 10 guys in the US, 5 in Europe and 2 in China." 

---

So though I understand we only have so many qualified witnesses, can only afford so many, voters can't vet too many, and at a certain level it decreases performance and even impacts security.  I'm definitely in favour of pushing it to the highest reasonable threshold.

Geographical diversity is also a powerful marketing tool for decentralisation if our witness number is limited.

If you can say BitShares is processed from 20-30 countries, including many that hate each other. Then you can sell a decentralised global company and exchange that is resistant to interference or attack despite technically having a small amount of witnesses.

 

[topcandle]

If Bitshares marketcap gets large enough, do you want to go back to 101 witnesses?  How large would that market cap have to be?

[jakub]

@bytemaster, could you let me know if I understand your line of thinking correctly:

(1) 17 is not so good as 25 or 51 in terms of preventing an attack (because of reasons raised by monsterer and Empirical1.2)
(2) But 17 is much easier to work with than 25 or 51 when we are in a recovery mode and need to coordinate quickly.

Is it the reason you prefer 17 to something higher?

I picked 17 to be the most controversial LOW number I could to stimulate the biggest possible debate... and because it is prime.

The test network currently has more than 17 unique nodes on it.    Like I said earlier I only want to set the pay rate high enough to make it a competition to get a slot.   I also want to reserve as much of the available budget for development and maintenance as possible

So it has nothing to do with the dilemma between being able to prevent an attack and being able to efficiently recover from one?

[bytemaster]

@bytemaster, could you let me know if I understand your line of thinking correctly:

(1) 17 is not so good as 25 or 51 in terms of preventing an attack (because of reasons raised by monsterer and Empirical1.2)
(2) But 17 is much easier to work with than 25 or 51 when we are in a recovery mode and need to coordinate quickly.

Is it the reason you prefer 17 to something higher?

I picked 17 to be the most controversial LOW number I could to stimulate the biggest possible debate... and because it is prime.

The test network currently has more than 17 unique nodes on it.    Like I said earlier I only want to set the pay rate high enough to make it a competition to get a slot.   I also want to reserve as much of the available budget for development and maintenance as possible

[jakub]

@bytemaster, could you let me know if I understand your line of thinking correctly:

(1) 17 is not so good as 25 or 51 in terms of preventing an attack (because of reasons raised by monsterer and Empirical1.2)
(2) But 17 is much easier to work with than 25 or 51 when we are in a recovery mode and need to coordinate quickly.

Is it the reason you prefer 17 to something higher?

[Riverhead]

That's where the low number of witnesses helps us: it's much easier to verify that 17 witnesses are unique than to verify that 51 witnesses are unique.

If they are anonymous, it's impossible to verify uniqueness.

Well, that's the thing, and to BM's point about choosing the witnesses. The location/servers/etc. can be in anonymous locations with anonymous IPs but the actual witnesses should at least be known to the Delegates.

Once the platform scales to beyond what can be run on a $100 dedicated server it's going to be pretty hard for a witness to stay hidden and still provide good service and confidence to the shareholders. Perhaps some mix of anonymous and known witnesses where 75% need to be known in a network of trust.

[donkeypong]

I think the core team of developers should focus on making bitshares/graphene a platform that's easy to build on.  End of the day, chains that have lots of 3rd party development (and thus utility), will win.

Bingo.

[monsterer]

That's where the low number of witnesses helps us: it's much easier to verify that 17 witnesses are unique than to verify that 51 witnesses are unique.

If they are anonymous, it's impossible to verify uniqueness.

[jakub]

I think it is highly unlikely that you could convince 8 people to collude without the other 8 finding out about it.
It would mean that 8 times in a row you would have to be lucky that the person you start convincing eventually joins your conspiracy without telling anyone about your plans.

And if those 8 people are really just one person?

That's where the low number of witnesses helps us: it's much easier to verify that 17 witnesses are unique than to verify that 51 witnesses are unique.

[monsterer]

I think it is highly unlikely that you could convince 8 people to collude without the other 8 finding out about it.
It would mean that 8 times in a row you would have to be lucky that the person you start convincing eventually joins your conspiracy without telling anyone about your plans.

And if those 8 people are really just one person?

[jakub]

If you're not targeting the altcoin market and want to break out into the real world - how will you convince banks and the establishment in general to put their trust and financing behind a system in which 8 'adversaries' working together can bring the entire thing to a grinding halt?

edit: and in the worst case at 0 cost

I think it is highly unlikely that you could convince 8 people to collude without the other 8 finding out about it.
It would mean that 8 times in a row you would have to be lucky that the person you start convincing eventually joins your conspiracy without telling anyone about your plans.

I'd guess it works like this: normally people trust those whom they consider experts in a given field. So our task is to have the support of "technical experts" that 17 is enough to make it safe.

The MaidSafe design requires just 4 copies of file chunks to store data safely on a decentralized network. At first I thought this number was way too small but then I read the experts' opinions and now I think 4 is OK if they so.