Topic: [XMR] Monero Improvement Technical Discussion - page 8. (Read 14717 times)

Shit, only skimmed this thread but this popped out at me. Is there that shit on the chain right now or are you warning of a possibility? Is anyone with the chain getting setup? To remove would take a hard fork prune?

How about adding a option to boycott specific transaction meta data and maintaining a blacklist? Is there another place where this is being discussed?

Cool, thanks for those links - will read up on them and try to understand them!!

re: the first bold, I agree, which is why these things would need to be done before things get too cemented. Furthermore, one of the pillars of Monero is decentralization - so I tend to believe that efforts would be made to maintain decentralization even at the cost of switching things as fundamental as the POW.

re: second bold - I agree to a degree ( snarf ), but its the "some degree" that matters, IMO. At the existing difficulty, it is still possible for your average PC consumer to mine monero, even though there are probably definitely botnets, AWS cowboys, and sysadmins already on the network. This is *not* the case with bitcoin and soon to be litecoin with the scrypt asic development.

re: third bold: e x a c t l y. thats what a theoretical polymorphism would do. A guarantted 10 year window where its guaranteed that the only thing that will work is a CPU.... granted, its impossible to know CPU architecture in the future.... but, we got math.

and there will soon be a missive detailing your point re: voting through forks, if I understood correctly.

The following approaches, being IO bound with a changing dataset achieve "some" extra level of polymorphism and although the PoW as it stands has strong merits for ASIC resistance, IMO they improve upon it somewhat.


http://boolberry.com/files/Block_Chain_Based_Proof_of_Work.pdf
https://github.com/ethereum/wiki/wiki/Ethash

For instance, in the case of the blockchain PoW (wild keccak) faster verification resulting in faster syncing. I believe last year someone conducted a test in an infographic for initial blockchain download and XMR was 200+ minutes, whereas BBR came in at about 13 minutes.  I'm trying to find where it was posted.

 I can't conceive a situation at the cryptonote protocol level where you could randomly switch algo's outright though. Voting through forks would be inelegant and cause much disruption.  Much like modifiying the emission curve, even if it was practical on a technical level, it would seem that it would far from ideal from an econ standpoint. You can't just switch what people signed up for retroactively.

  Also, you cannot ever stop the tendency to drift towards mining centralization- even the classic notion of 1 CPU, 1 Vote can be abused by sysadmins, botnet owners etc, AWS cowboys. I know this well as I was mining BTC before there were optimized CPU miners- and no GPU miners nor pools, and it was not even called mining, but "Generating'. It was always centralized to some degree.  With the rolling snowball effect of those who are profiting ploughing in more towards expansion you can only ever attempt to stave off specialized custom silicon.

I wonder if there's a way to make morphing POW code, such that every n years the POW function gets modified.

This would prevent mining centralization, and do so in a way that avoids politics, because changing the POW would definitely cause an uproar from the mining camp.

In the case that ASICS *are* developed for Monero, this would disrupt that significantly.
If the only thing that exists in 10 years are GPU operations, then it would only disrupt it in terms of software development.
if the only thing is CPU mining, then all it would do is cement further CPU mining.

I would support the development of a MoneroAsset daughter chain, especially if it enabled refining the primary currency chain. Removing the ability to use the primary currency chain as more than what its supposed to be - accounting - would be very beneficial, IMO. Especially because Monero already has to deal with an inherently larger blockchain with interesting hurdles for pruning. Having separate chains would also make it explicitly simpler to differentiate between the everyday user (someone who runs a currency-only node) vs. enthusiasts or vested interests running multiple chains. Of course, incentivizing maintenance of the daughter chain is a different beast, unless I'm not understanding daughter chains properly.

Yeah, generally, I feel that trying to multipurpose the ledger is something to be avoided - especially in the case of Monero where the data is cryptographically obscured. I don't think every blockchain should be viewed as an oppurtunity for data storage - "metadata normally stuffed into the mainchain". The primary currency chain should be just that - a ledger where it is indicated that I sent n things to someone else. Keep it simple stupid.

and yes, preventing kiddie porn on the blockchain is always a good thing.

damnit I wish I knew how to code so I could get the data from the blockchain to try and come up with a formula for the floating minimum fee adjustment proposal outlined above. Gah. It'd be fun to make pretty graphs.

I would 100% support dropping tx_extra *if* we have MoneroAsset etc. duaghter-chains to deal with the metadata normally stuffed into mainchain. Having a separate optional output (or tx) identifier is a good replacement.

Removing support for payment IDs entirely could be a consensus change. I have no idea if they intend that.

The transaction format could be made a lot cleaner if tx_extra were remove and a fixed size transaction-key field were added instead. Plus it would remove one obvious way for scumbags to stuff kiddie porn on the blockchain.

Yes indeed. Sorry, I didn't really mean to come across as snarky (not sure if I did).

Neither of these change any consensus rules in any way, they're just attempts to recommend standards for recipients to differentiate payments.

So one viewkey for multiple spendkeys, is that what bytecoin is doing ? I am more favorable to this option rather than payment_id

The idea is to have one view key with multiple spend keys. That gives you multiple public keys (addresses) where half is common and half is unique. This allows scanning for all of these addresses much more quickly than scanning for completely different addresses (for the use case of an exchange or web wallet where each user has his own address -- instead of the current practice of one exchange address plus payment IDs).

The new Bytecoin "breakthrough", where they're "depreciating" payment IDs. It is a potential solution, and I'd be really interested in seeing an actual cost analysis instead of generalities.

Obligatory reading, how Monero identified some weakness and proposed changes to betterment of the cryptonote protocol:

https://lab.getmonero.org/pubs/MRL-0001.pdf
https://lab.getmonero.org/pubs/MRL-0002.pdf
https://lab.getmonero.org/pubs/MRL-0003.pdf
https://lab.getmonero.org/pubs/MRL-0004.pdf

I don't understand .....'

good idea on the thread title change.

So pretty. I haven't any useful input on fees.

What do people think about serialized stealth payment IDs versus unique spend keys with single viewkey?

Edit: try to keep up!

Edit2: maybe you should change title to "[XMR] Monero Improvement Technical Discussion" or something.

Lots of posts deleted because they were not technical.

I grow weary of the constant trolling, inverse trolling, talk of "investing", nonsense etc. This is cool software / technology that happens to be explicitly tied to money.

This thread is about Monero technical ideas, discussion, ONLY. If you barge in here with something non-technical, I will delete it. This is not a support thread. I will point you in the right direction, then delete your post.

Yes, you might say "there's a monero specific forum at forum.getmonero.org". Yes there is. People don't use it for their own reasons.

My first post will be a copy of the discussion smooth and I were having regarding trying to figure out how to automodify the minimum fees. I hope that this cleaner thread will spur more actual discussion as opposed to the neverending "woh look market cap! hey I have a scam detector! who wants some spam?"