Agreed - I am a little confused now also what the real difference is (apart from perhaps trying to add an extra "check").

If it is then I'd suggest that a request be made for them to "warn before continuing" if the tx fee is greater than the amount being sent to other addresses.

EDIT: In fact the warning should probably be there if the fee is much greater than *what is needed* (although hard to say they are *responsible* for this loss I would hope they'd fix the issue and refund this guy).

Yes - but only when you *send* (where that address is one of the UTXOs). If you create a new address and "receive" BTC then your public key has *not hit the blockchain* (and won't until you spend that UTXO).

Note also that I am pretty sure that if you sign a "verifymessage" then your public key is included in that "signature" (so don't sign messages with a bitcoin address that you haven't "spent from" unless you really have to).

The problem is if someone ever finds a weakness in ECDSA that might enable a private key to be recovered from a public one. Although in general this is a "theoretical" problem a "real" problem was found (last year from memory) due to the use of "non-random" K values (you'll need to google that to find out the gory details).

Provided an address has not been used then its public key does not appear in the blockchain (only the hashed "address") so it keeps your "savings safer". Of course if a weakness was found in ECDSA that did allow "private keys to be worked out" then you would not be able to "spend your BTC" until some kind of "fix" was made available (as your first attempt to do so would publish the public key making it vulnerable to a tx "replacement" that might steal your BTC).

Maybe so - but we are talking about people "typing in from what they read" (so mistakes are bound to happen).

The fact that if you create any bad "brainwallet" it will have its funds exhausted within minutes suggests that people are happy to run such bots in the hope of finding BTC (e.g. the "horse staple battery" address).

Aha - *but* it may be possible using vanitygen to come up with another address starting with 1ciyam3htJ in the next block (depending upon what GPU power you have with regards to vanitygen). To do the same with the "checksum", however, would be *much, much harder* (as you'd have to try many more addresses to get such a match).

Thanks @BurtW - although I am sure this whole idea *grates* on you (as I guess it does with many others) I think it is valuable to at least work out the best practice if one is going to "publish an address" (especially on a "billboard" or the like).

Let's see if we can clarify it a bit (I might not have correctly understood the algo).

If we have:

1ciyam:2adfd then that is currently an okay MinAddress (assuming such an address exists in that block)

but if we have:

1ciyam:2adfe then "couldn't this also be a correct MinAddress" (if a new 1ciyam address had appeared in the next block)?

Assuming that this assumption is correct then:

1ciyam:2adfeTeju would be invalid as the Teju would *not match with the new 1ciyam address*.

(whereas 1ciyam:2adfdTeju would be valid as Teju does match the checksum of the 1ciyam address in the previous block)

EDIT: Or maybe there simply can't be two 1ciyam's in which case there is no such issue (part of the problem is that I am assuming 1ciyam rather than 1ci as I want to keep the "vanity address" prefix but maybe this is also *irrelevant*).

But as the last x characters of a BTC address *are* a checksum (for the rest of the address) then why not use that rather than create another checksum?

I think the issue that was been raised was that of a "typo".

If one types 1ciyam:2adfe rather than 1ciyam:2adfd then presumably you could end up with an incorrect address.

If all or part of the checksum was included then 1ciyam:2adfdeJu would be "invalid" (unless it was actually the same address).

I don't think it could work like that (otherwise you'd break copy and paste per my sig example) - I think that the symbol (being a : or something else) itself would imply that the 4 characters *before* the : are the last 4 from the address.

so: 1ciyamTeju:2adfd

(but is now starting to not be so "minimal")

Also I guess you could *repeat* the checksum characters like this:

1ciyam3htJit1feGa26p2wQ4aw6KFTejU:2adfdTejU

It is how "to find" said vanity address that this addresses (not creating it).

Think you must have missed this:


Already thunk it!

But glad to see you agree with my suggestion.

Thanks - note that it is somewhat a "marketing issue" also (so I was trying to make it look nice while still working with good old copy and paste - including touch screens).

Your "edit" idea perhaps makes sense for the "web site widget" but won't work on a "billboard" (unless it is a digital "smart billboard" I guess).

Personally I would create a small wallet for such "tips" or use "coin control" to get rid of some small amount (perhaps another tip I got myself) but of course I get your point that for most people this would be "inconvenient" or "too difficult".

I think that both your and Danny's point with this is really that "publishing an address" is a bad idea full stop (as even if that address "changes with each tip/donation" someone can just monitor the website to keep a track of all such addresses that were given out taking you back to square #1).

In favour of the "short address" idea I have had at least one or two experiences in "buying BTC for cash" when "I did not have my computer handy" - in that case having either a "known public address" (or being able to find one easily with a "short name") was actually "the only way I was going to be able do the deal".

Seriously? Do you think that someone needs to go to someone's website or contact them in order to just give a tip?

I can understand if you are offering a "service" that it would make sense but if you are just tipping someone for a great post then I do not see the harm at all.

1ciyam3htJit1feGa26p2wQ4aw6KFTejU:2adfd

Hmm... you could put part or all of the normal address checksum in the MinAddress like the above.

@DannyHamilton

You seem to have missed the "tip address" argument to "address re-use" - if people had to go to a website in order to get a unique address to "send a tip" then very few people would ever actually get a tip (and I have had a few sent to my 1ciyam address over the years).

Another variation of the "tip address" is the "billboard sign" one (yes you could also put a QR code there but if the person doesn't have a camera then with this they can still "jot down the address" to find the full address later).

Let me know when it is added and I'll even change the address in my sig to include it.

1ciyam3htJit1feGa26p2wQ4aw6KFTejU:2adfd

Nicer but @ might get confused with email - perhaps $ instead?

(or if $ is going to be controversial then ! or : would be okay)

1ciyam3htJit1feGa26p2wQ4aw6KFTejU!2adfd
1ciyam3htJit1feGa26p2wQ4aw6KFTejU:2adfd