Quote from: DooMAD on November 17, 2018, 07:59:50 AM

Quote from: cfbtcman on November 17, 2018, 03:25:57 AM

its sufficient that in a group of 1000 nodes, 500 save backup of 50% of the data (Type A node) and the other 50% can save backup of the other 50% of the data (Type B node), its a waste of space all the nodes in the network save all information repeated thousands of times.

Then each node Type A can "ask" to other node Type B the information is trying to find and vice-versa and gets the information anyway

Bitcoin was designed to be trustless. The idea of running a node is that you can validate and verify every single transaction yourself. If you run a Type A node, you would have to trust the Type B nodes to do half of the validation for you. If you're going to do that, why not just trust Visa and forget all about Bitcoin?

finally doomad sees the light about "compatibility not = full node".. and how "compatible" is not good for the network..
one merit earned... may he accept the merit and drop that social drama debate now he seen the light.

onto the topic
the hard fork of removing full nodes that can only accept 1mb blocks has been done already, in mid 2017. reference the "compatible" nodes still on the network are not full nodes no more.

all is required is to remove the "witness scale factor" and the full 4mb can be utilised by legacy transactions AND segwit transactions.
this too can have positives by removing alot of wishy washy lines of code too and bring things back inline with a code base that resembled pre segwit block structuring to rsemble a single block structure where everything is together that doesnt need to be stripped/"compatible".
yes the "compatible" nodes would stall out and not add blocks to their hard drive. but these nodes are not full nodes anyway so people using them might aswell just use litewallets and bloom filter transaction information they NEED for personal use.

we will then have the network able to actually handle more tx/s at a 4x level rather than a 1.3-2.5x level(which current segwit blockstructure LIMITS (yep even with 4mb suggested weight. actual calculations limit it to 2.5x compared to legacy 1mb))

...
as for how to scale onchain.
please do not throw out "LN is the solution" or "servers will be needed" or "you cant buy coffee"
1. instead of needing LN for coffee by channelling to starbucks. just onchain use a tx to btc buy a $40 starbucks giftcard once a fortnight.

after all from a non technical real life utility perception of average joe. if your LN locking funds with starbucks for a fortnight anyway. its the same 'feel' as just pre-buying a fortnights worth of coffee via a giftcard.
(it also solves the routing, multiple channel requirement for good connection, spendability, also the other problems LN has)

2. onchain scaling is not about just raising the blocksize. its also about reducing the sigops limit so that someone cant bloat/maliciously fill a block with just 5 transactions to use up the limits. EG blocksigops=80k and txsigops=16k meaning 5 tx's can fill a block should they wish. this is by far a bad thing to let continue to be allowed as a network rule.*

3. point 2 had been allowed to let exchanges batch transactions into single transactions of more in/outputs so that exchanges could get cheaper fee's. yet if an exchange is being allowed to bloat a block alone. then that exchange should be paying more for that privelige, not less.
(this stubbornly opens up the debate of should bitcoins blockchain be only used by reserve hoarders of multiple users in permissioned services(exchanges/ln factories).. or should the network be allowing individuals wanting permissionless transacting).. in my view permissioned services should be charged more than an individual

4. and as we move away from centralised exchanges that hoard coins we will have less need for xchanges to batch such huge transactions and so there will be less need of such bloated transactions

5. scaling onchain is not just about raising the blocksize. its about making it more expensive for users who transact more often than those who transact less frequently.
EG imagine a person spend funds to himself every block. and was doing it via 2000 separate transactions a block (spam attack)
he is punishing EVERYONE else. as others that only spends once a month are finding that the fee is higher, even though they have not done nothing wrong.
the blocks are still only collating the same 2000tx average. so from a technical prospective are not causing any more 'processing cost' to mining pool nodes tx's into block collation mechanism. (they still only collate ~2000tx so no cost difference)
so why is the whole network being punished. due to one persons spam.

the person spending every block should pay more for spending funds that have less confirms than others. in short the more confirms your UTXO has the cheaper the transactions get. that way spammers are punished more.
this can go a stage further that the child fee also increases not just on how young the parent is but also the grandparent

in short bring back a fee priority mechanism. but one that concentrates on age of utxo rather than value of utxo(which old one was)