Let's say that JimBob came along offering a lollipop to everyone for every $1000 they could prove they held in a US bank on a particular day. (The BCH network is offering credits to everyone on the BTC network in proportion to how much BTC they can prove they had on a particular day.)
You, as a customer, tell your bank, "yes, confirm to JimBob that I had $30,000 in my account on that day!" (You provide notarized authorization, etc.) The bank says no. Would that fly? Maybe.
Instead, they call up JimBob and say, "Look, we have $3 million in our bank vault. Give us our 3,000 lollipops."
Would that fly? I don't think so. (Though Coinbase hasn't done that last part ... yet.)
Well, really, the problem is that JimBob is in total control of the interaction here. Maybe your bank says, "Sure, here's an affidavit showing you had $30,000 on date X", but that's not good enough for JimBob; he needs to see the actual, physical bills and needs proof that _those specific bills_ were in your bank's vault at the time.
This process puts a fair amount of burden on the "bank", because they need to deal with JimBob's byzantine lollipop claims process _directly_ with their own cash, rather than issuing a standard affidavit that'll work equally well for MarySue's pop-tart claims process. Or, in more technical terms, _every time_ anyone comes up with a new coin, coinbase would have to dust off the cold vault, generate the proofs of ownership, reassign them to the new owners, then move the bitcoin over into newly generated cold addresses to restore the security posture of the cold vault coins.
Moreover, creating new types of altcoins is a process that can be easily automated, and at little cost to whoever is instigating this creation process. Of course, most of these coins won't ever gain any popularity or value - but then what's the threshold? At what level of popularity does a broker need to do anything? What happens if a coin is totally unpopular initially but gains popularity later - do brokers need to backfill? What if the key material has been rotated away and destroyed in the meantime?
For that matter, how do we measure popularity? Volume? It's easy to spam volume if you control the first exchange for your altcoin. Market cap? It's easy to fake market cap and unrealized gains when nobody can transfer coins yet. Hashrate? Who knows when that'll settle down. In any case, it's not the sort of thing that can really be objectively determined too quickly after a split.
These don't become an issue with normal securities because, in normal stock market securities there is only one authorized issuer who is able to issue splits, and this issuer incurs real, significant administrative costs for doing so. They also must involve credentialed third parties, and all participants face regulatory scrutiny, fines, or even jail time if they deliberately set out to abuse the system.
Compare to the creation of an altcoin, where the costs of creating a new altcoin are tiny, but the costs of dealing with the fallout by brokers can be quite large. And, of course, there is effectively zero regulation that would give penalties to deliberate abusers. There's therefore a large cost and risk imbalance between altcoin creators and brokers who have to deal with the new altcoins. This sort of cost/risk imbalance results in, effectively, a sort of denial of service threat against bitcoin brokers if it's allowed to stand.
For this reason, the precedent set by this split is going to be of great importance in the future of bitcoin and altcoin brokers - if brokers must deal with splits, it will mean that they will need to segregate coins and keep them online ("hot") at all times in order to be able to deal with any split immediately and/or retroactively. This in turn will hurt the security posture of these currencies substantially. The risk of screwing a split up will also hang over any service that holds *-coins on behalf of their customers or other third parties.
This will also impact the legal feasibility of off-chain transaction aggregation schemes (e.g. lightning) not feasible, as there's no guarantee that the next harebrained altcoin split will be able to deal with redeeming the complex transaction scripts involved with e.g. an unsettled lightning channel - and if it doesn't, who is liable?
So, in short, _requiring_ brokerages to deal with altcoin splits and redemptions opens up a huge legal can of worms that is probably best left closed, for the viability of the overall altcoin landscape. Yes, it sucks that coinbase didn't do something for this particularly public split, but if they did they'd have to be processing splits every other day, or perhaps even more frequently.
Disclaimer: I have no position, long or short, in bitcoin or any other similar "altcoin" assets, nor do I have any plans to open such a position in the forseeable future. However, I do find it fascinating to watch this legal/economic experiment evolve from a safe distance.
That's a lot of text that dances around the problem. Wouldn't the issue be solved if exchanges just offer IOUs for a particular chain instead of claiming to be a "wallet" or "bank" holding your asset? If they want to claim they are holding a user's "coins" (which is a misnomer to be honest and is an inaccurate representation of how blockchains work), then they better have a particular address and private key somewhere for ever user that holds assets on their exchange.
Well, the problem is that this means exchanges need to actually implement support for every chain. If I tomorrow go muck around with my bitcoin node and make it fork, why shouldn't they be required to issue you IOUs for that altcoin as well, after all? What if they didn't know that a fork happened, went and rotated all their cold storage keys, destroyed the old private keys, and later you ask for your coins on the secret chain?
You have to draw the line somewhere, and, as a practical concern, determining that a fork meets this line retroactively adds a great deal of complexity and risk to the process. Further, as we've seen here, even a short delay in executing the split raises a great deal of ire and threats of lawsuits.
So, as I see it, the only consistent positions to take would be to either:
1. Not require split tracking at all. People who want to enjoy the benefits of such a split can simply move their bitcoins to a wallet under their direct control, or to an exchange who has agreed to track the split. [what happened here - a market-driven solution, essentially]
2. To legally mandate split tracking if proponents of the split agree to make a market - that is, agree to buy at a particular price for a particular period of time - thus ensuring that the coins have value [unlikely to ever come into play due to the enormous cost of such a venture]
3. To legally mandate split tracking for absolutely every altcoin, no matter how awkward or technically difficult it may be to do so.
Keep in mind that implementing 3 means the exchange might need to run software created by the altcoin's promoters - because keeping up with your own software for every altcoin would take too much development resources - and would be exposed to the exchange's private keys in order to perform hot wallet transactions that the altcoin's network will accept. Thus, if exchanges are forced to run software from any random Joe's fork, this would present a real security risk that backdoored software might be slipped in - this means that they would be forced to switch to split wallets by customer so a single customer using JoeCoin won't compromise funds held by other customers, but having this split out makes maintaining cold wallets impractical.
> Well, the problem is that this means exchanges need to actually implement support for every chain
No, read my comment again, carefully. If they are providing IOUs for BTC, that means there is no private key associated with the debt. If say for instance they owe a user 10 BTC, they could give them any 10 BTC. if a new chain pops up, it doesn't matter, because the contract is explicitly an IOU for 10 BTC, not BCH or BCC.
the problem arises when the exchange implies they are holding "your coins" for you like a safe deposit box, which means there's a private key associated with the coins, and thus that key would work with any chain splits, including BCH/BCC.
just curious, do you understand how blockchains work? i.e. private keys and the transaction protocol? if you weren't familiar with it, then that would explain why you are having a hard time understanding my comment.
I think I misread your comment - I had interpreted it as suggesting that exchanges should be required to issue IOUs for new forked assets, which has serious practical issues.
Of course, saying that coinbase balances are IOUs for bitcoins is basically what happened here, and apparently people aren't happy about it. While you could change the marketing material, I don't think it would do much to change the outrage in this case; I very much doubt people will pay attention to such a nuance when they apparently couldn't be bothered to temporarily withdraw their coins.
Then how about this: JimBob constructed a perfect copy of everyone's bank safety deposit box key and its contents, including the keyhole[1]. The contents of the copy are free for the taking if you can open it with the key.
Your bank happens to have an access control policy where they hold the safety deposit box key for you and you have to pick it up from them each time (after authenticating).
You ask the bank for your key so can go to the JimBob version and get your copied stuff. They refuse because of a security policy about taking keys offsite.
You offer to close the box account and take your (original) stuff out so that there will be no security compromise. They refuse because that would compromise the box and cost them a usable one.
You offer to pay for the cost of replacing the box with an uncompromised one. They still refuse.
That is the position that Coinbase is in. (Users haven't offered to pay for the replacement of the private keys that coinbase uses, but this is a trivial cost, and establishes the barrier that prevents coinbase from facing unbounded costs to deal with every fork.)
EDIT: And for an analogy that favors "Coinbase doesn't owe you anything": Assume we're under a gold standard. Assume that some goldsmiths just give you regular dollar bills as your warehouse receipt "since they're redeemable with the government banks anyway". Most people don't care much about the difference, but each goldsmith keeps a list of the serial numbers they gave out so they don't have to honor arbitrary demands.
Then (of course) the government goes off the gold standard. But some goldsmiths are like, "no, that's BS, we will still honor your claims, just bring us a dollar bill that's on our serial number list". So there's a mad dash to get dollar bills that were used as goldsmith receipts, and some people ask for their deposits back from their bank, but the bank (like Coinbase) insists that they only have to give you banknotes of their choice, not the original ones you deposited, which was all they ever promised.
(Modified scenario: you kept the dollar bills from the goldsmith in a safety deposit box, and the bank replaces them with new dollars and never redeems them for gold.)
[1] You can further assume that the copy is some debased, lower value version of arbitrary percentage and it doesn't affect the analogy.
Coinbase is more similar to a bank than it is to a wallet.