[bitcoin-dev] BIP75 - Out of Band Address Exchange

Discussion:

James MacWhyte via bitcoin-dev

2016-03-10 21:43:29 UTC

Hi everyone,

Our BIP (officially proposed on March 1) has tentatively been assigned
number 75. Also, the title has been changed to "Out of Band Address
Exchange using Payment Protocol Encryption" to be more accurate.

We thought it would be good to take this opportunity to add some optional
fields to the BIP70 paymentDetails message. The new fields are:
subtractable fee (give permission to the sender to use some of the
requested amount towards the transaction fee), fee per kb (the minimum fee
required to be accepted as zeroconf), and replace by fee (whether or not a
transaction with the RBF flag will be accepted with zeroconf). I know it
doesn't make much sense for merchants to accept RBF with zeroconf, so that
last one might be used more to explicitly refuse RBF transactions (and
allow the automation of choosing a setting based on who you are transacting
with).

I see BIP75 as a general modernization of BIP70, so I think it should be
fine to include these extensions in the new BIP, even though these fields
are not specific to the features we are proposing. Please take a look at
the relevant section and let me know if anyone has any concerns:
https://github.com/techguy613/bips/blob/master/bip-0075.mediawiki#Extending_BIP70_PaymentDetails

The BIP70 extensions page in our fork has also been updated.

Thanks!

James

Andreas Schildbach via bitcoin-dev

2016-03-11 11:54:52 UTC

Permalink

I think it's a bad idea to pollute the original idea of this BIP with
other extensions. Other extensions should go to separate BIPs,
especially since methods to clarify the fee have nothing to do with
secure and authenticated bi-directional BIP70 communication.

Post by James MacWhyte via bitcoin-dev
Hi everyone,
Our BIP (officially proposed on March 1) has tentatively been assigned
number 75. Also, the title has been changed to "Out of Band Address
Exchange using Payment Protocol Encryption" to be more accurate.
We thought it would be good to take this opportunity to add some
subtractable fee (give permission to the sender to use some of the
requested amount towards the transaction fee), fee per kb (the minimum
fee required to be accepted as zeroconf), and replace by fee (whether or
not a transaction with the RBF flag will be accepted with zeroconf). I
know it doesn't make much sense for merchants to accept RBF with
zeroconf, so that last one might be used more to explicitly refuse RBF
transactions (and allow the automation of choosing a setting based on
who you are transacting with).
I see BIP75 as a general modernization of BIP70, so I think it should be
fine to include these extensions in the new BIP, even though these
fields are not specific to the features we are proposing. Please take a
https://github.com/techguy613/bips/blob/master/bip-0075.mediawiki#Extending_BIP70_PaymentDetails
The BIP70 extensions page in our fork has also been updated.
Thanks!
James
_______________________________________________
bitcoin-dev mailing list
https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev

Justin Newton via bitcoin-dev

2016-03-11 22:43:48 UTC

Permalink

I think we would be open to either leaving them in, or doing a separate
BIP. What do others think? Iâd prefer to keep them together if the
changes are non-controversial just to cut down on #of BIPâs, but thats not
a strong preference.

On Fri, Mar 11, 2016 at 3:54 AM, Andreas Schildbach via bitcoin-dev <

Post by Andreas Schildbach via bitcoin-dev
I think it's a bad idea to pollute the original idea of this BIP with
other extensions. Other extensions should go to separate BIPs,
especially since methods to clarify the fee have nothing to do with
secure and authenticated bi-directional BIP70 communication.

https://github.com/techguy613/bips/blob/master/bip-0075.mediawiki#Extending_BIP70_PaymentDetails

Post by James MacWhyte via bitcoin-dev
The BIP70 extensions page in our fork has also been updated.
Thanks!
James
_______________________________________________
bitcoin-dev mailing list
https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev

_______________________________________________
bitcoin-dev mailing list
https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev

--
Justin W. Newton
Founder/CEO
Netki, Inc.

***@netki.com
+1.818.261.4248

Andreas Schildbach via bitcoin-dev

2016-03-12 15:00:17 UTC

Permalink

Replying to the "fee" part of BIP75 (which as already noted should go to
a different BIP number imho):

It makes to sense to let the payee define a fee *rate*. The payee
doesn't know anything about how the payer's wallet is structured. In
extreme cases, as a payer I would keep all my tiny UTXOs (which would be
unspendable in a economic way) for the one payee who is willing to pay a
high enough rate...

Rather, I propose an absolute amount that the payee is willing to cover
should be declared.

Also, in order to avoid disputes I suggest the amount should be deducted
from the BIP70 payment message amount already. A wallet which
understands BIP75fee would add these two up for *display* puposes only.
The wallet should continue to use the existing fee policies. If it can
send the amount as specified by BIP70 and the fee is below the BIP75fee
amount, it would not mention any fees to the user. If it exceeds, it
would display just the exceeding amount.

Post by Justin Newton via bitcoin-dev
I think we would be open to either leaving them in, or doing a separate
BIP. What do others think? I’d prefer to keep them together if the
changes are non-controversial just to cut down on #of BIP’s, but thats
not a strong preference.
On Fri, Mar 11, 2016 at 3:54 AM, Andreas Schildbach via bitcoin-dev
I think it's a bad idea to pollute the original idea of this BIP with
other extensions. Other extensions should go to separate BIPs,
especially since methods to clarify the fee have nothing to do with
secure and authenticated bi-directional BIP70 communication.

(whether or

Post by James MacWhyte via bitcoin-dev
not a transaction with the RBF flag will be accepted with zeroconf). I
know it doesn't make much sense for merchants to accept RBF with
zeroconf, so that last one might be used more to explicitly refuse RBF
transactions (and allow the automation of choosing a setting based on
who you are transacting with).
I see BIP75 as a general modernization of BIP70, so I think it

should be

Post by James MacWhyte via bitcoin-dev
fine to include these extensions in the new BIP, even though these
fields are not specific to the features we are proposing. Please

take a

Post by James MacWhyte via bitcoin-dev
look at the relevant section and let me know if anyone has any

https://github.com/techguy613/bips/blob/master/bip-0075.mediawiki#Extending_BIP70_PaymentDetails

_______________________________________________
bitcoin-dev mailing list
https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev
--
Justin W. Newton
Founder/CEO
Netki, Inc.
+1.818.261.4248 <tel:+1.818.261.4248>
_______________________________________________
bitcoin-dev mailing list
https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev

James MacWhyte via bitcoin-dev

2016-03-17 01:23:09 UTC

Permalink

We have removed the BIP70 field extensions from this BIP and will save that
for another time. A PR to add our documentation to the main repo has been
submitted.

James

On Sat, Mar 12, 2016 at 8:36 AM Andreas Schildbach via bitcoin-dev <

Post by Andreas Schildbach via bitcoin-dev
Replying to the "fee" part of BIP75 (which as already noted should go to
It makes to sense to let the payee define a fee *rate*. The payee
doesn't know anything about how the payer's wallet is structured. In
extreme cases, as a payer I would keep all my tiny UTXOs (which would be
unspendable in a economic way) for the one payee who is willing to pay a
high enough rate...
Rather, I propose an absolute amount that the payee is willing to cover
should be declared.
Also, in order to avoid disputes I suggest the amount should be deducted
from the BIP70 payment message amount already. A wallet which
understands BIP75fee would add these two up for *display* puposes only.
The wallet should continue to use the existing fee policies. If it can
send the amount as specified by BIP70 and the fee is below the BIP75fee
amount, it would not mention any fees to the user. If it exceeds, it
would display just the exceeding amount.

Post by Justin Newton via bitcoin-dev
I think we would be open to either leaving them in, or doing a separate
BIP. What do others think? Iâd prefer to keep them together if the
changes are non-controversial just to cut down on #of BIPâs, but thats
not a strong preference.
On Fri, Mar 11, 2016 at 3:54 AM, Andreas Schildbach via bitcoin-dev
I think it's a bad idea to pollute the original idea of this BIP with
other extensions. Other extensions should go to separate BIPs,
especially since methods to clarify the fee have nothing to do with
secure and authenticated bi-directional BIP70 communication.

Post by James MacWhyte via bitcoin-dev
Hi everyone,
Our BIP (officially proposed on March 1) has tentatively been

assigned

Post by Justin Newton via bitcoin-dev

Post by James MacWhyte via bitcoin-dev
number 75. Also, the title has been changed to "Out of Band Address
Exchange using Payment Protocol Encryption" to be more accurate.
We thought it would be good to take this opportunity to add some
optional fields to the BIP70 paymentDetails message. The new
subtractable fee (give permission to the sender to use some of the
requested amount towards the transaction fee), fee per kb (the

minimum

Post by Justin Newton via bitcoin-dev

Post by James MacWhyte via bitcoin-dev
fee required to be accepted as zeroconf), and replace by fee

(whether or

Post by James MacWhyte via bitcoin-dev
not a transaction with the RBF flag will be accepted with

zeroconf). I

Post by Justin Newton via bitcoin-dev

Post by James MacWhyte via bitcoin-dev
know it doesn't make much sense for merchants to accept RBF with
zeroconf, so that last one might be used more to explicitly refuse

RBF

Post by Justin Newton via bitcoin-dev

Post by James MacWhyte via bitcoin-dev
transactions (and allow the automation of choosing a setting based

Post by Justin Newton via bitcoin-dev

Post by James MacWhyte via bitcoin-dev
who you are transacting with).
I see BIP75 as a general modernization of BIP70, so I think it

should be

Post by James MacWhyte via bitcoin-dev
fine to include these extensions in the new BIP, even though these
fields are not specific to the features we are proposing. Please

take a

Post by James MacWhyte via bitcoin-dev
look at the relevant section and let me know if anyone has any

https://github.com/techguy613/bips/blob/master/bip-0075.mediawiki#Extending_BIP70_PaymentDetails

Post by Justin Newton via bitcoin-dev

_______________________________________________
bitcoin-dev mailing list
https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev

James MacWhyte via bitcoin-dev

2016-03-11 19:32:12 UTC

Permalink

That's a valid point, and one we had thought of, which is why I wanted to
get everyone's opinion. I agree the proposed field extensions have nothing
to do with encryption, but does it make sense to propose a completely
separate BIP for such a small thing? If that is the accepted way to go, we
can split it into two and make a separate proposal.

On Fri, Mar 11, 2016 at 5:48 AM Andreas Schildbach via bitcoin-dev <

https://github.com/techguy613/bips/blob/master/bip-0075.mediawiki#Extending_BIP70_PaymentDetails

_______________________________________________
bitcoin-dev mailing list
https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev

Andreas Schildbach via bitcoin-dev

2016-03-12 14:40:11 UTC

Permalink

Yes, it makes sense. A BIP is something people refer to, either just by
its number or by URL, and with multiple orthogonal "sub-BIPs" it's
difficult to refer to. We have this problem with BIP32 already -- all HD
wallets implement the derivation part of BIP32 but almost none do
implement the hierarchy part (and use BIP43/44 instead). I tried to
split up BIP32 into two BIPs later (without any content changes), but it
was declined because of its final state.

There is no harm in using a BIP only for a small thing, BIP numbers are
infinite.

Post by James MacWhyte via bitcoin-dev
That's a valid point, and one we had thought of, which is why I wanted
to get everyone's opinion. I agree the proposed field extensions have
nothing to do with encryption, but does it make sense to propose a
completely separate BIP for such a small thing? If that is the accepted
way to go, we can split it into two and make a separate proposal.
On Fri, Mar 11, 2016 at 5:48 AM Andreas Schildbach via bitcoin-dev
I think it's a bad idea to pollute the original idea of this BIP with
other extensions. Other extensions should go to separate BIPs,
especially since methods to clarify the fee have nothing to do with
secure and authenticated bi-directional BIP70 communication.

(whether or

should be

Post by James MacWhyte via bitcoin-dev
fine to include these extensions in the new BIP, even though these
fields are not specific to the features we are proposing. Please

take a

Post by James MacWhyte via bitcoin-dev
look at the relevant section and let me know if anyone has any

https://github.com/techguy613/bips/blob/master/bip-0075.mediawiki#Extending_BIP70_PaymentDetails

_______________________________________________
bitcoin-dev mailing list
https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev
_______________________________________________
bitcoin-dev mailing list
https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev