BIP43: Block usage 0 because it's already used in BIP32. #94

pull schildbach wants to merge 1 commits into bitcoin:master from schildbach:bip43-purpose-0 changing 1 files +3 −0

schildbach commented at 8:21 AM on August 16, 2014: contributor

It would be unfortunate if BIP32 wallets would at one day overlap BIP43 wallets.
voisine commented at 5:34 PM on August 16, 2014: contributor

I like it. I was going to write a BIP for the default BIP32 use case, but it would break the convention of using the BIP number as the purpose number. breadwallet and I believe also Hive are the projects using 0'
dabura667 commented at 6:44 PM on August 16, 2014: none

-1

BIP number as the purpose number

This was the thing I understood least.

So any BIP32 chain that uses BIP43 / 44 with BIP39 as the seed could decide to use 32' 39' 43' or 44'?

imo, BIP32 doesn't really specify "this is what you must use as your hierarchy" So setting a specification off to the side for it completely defeats the purpose of BIP 43/44.

"We make this BIP to finally decide on a HD wallet structure methodology, because BIP32 decided not to... but it kinda did... but that was not sufficient, so we made a new BIP... but we won't improve the old one, just... kinda do our own thing over here then... no one use m/0'/0/k,,, that's BIP32's inferior hierarchy."

I view BIP32 as "Here's how to make HD wallets" and BIP43/44 as "Here's how to order them"

Yet this PR seems to insinuate that A. BIP32 has a set in stone requirement of an HD structure. and B. That this structure is completely valid, and should be validated by BIP43 as a side-note.

I understand some wallets implemented m\0'\0\k already. But that was because there was no specifications.

Reinforcing the "use cases" as "This is how BIP32 Hierarchies should be done!" and then saying in your own BIP "HD wallets needs defined purposes!" seems very contradictory to me.
schildbach commented at 7:01 PM on August 16, 2014: contributor

@dabura667 The text you're quoting means the BIP number of the BIP implementing a BIP43 hierarchy. Currently, the only (published) BIP implementing BIP43 is BIP44.

BIP32 in the general case (using all acounts) is overlapping BIP43, so its incompatible. However, when you just use the default account (number 0), there are no overlaps as long as no one implements a BIP43 hierarchy with purpose 0. To prevent that from happening, I'm proposing this PR. All I want to say is "please don't use purpose 0".
schildbach commented at 7:04 PM on August 16, 2014: contributor

@dabura667 I agree with you we should remove as much as possible of BIP32-hierarchy. I tried some weeks ago and got plenty of disagreements. See #61
dabura667 commented at 7:15 PM on August 16, 2014: none

It sounds to me like they were waiting in a deprecation notice commit instead.

I personally see no need to remove the whole use case section in BIP32, just add an adendum to it saying "BIP43 renders this as obsolete, so don't use it."

Also it seemed like everyone saw your point... The only problem was inactivity and house cleaning.

That issue should be reopened and repurposed as an adendum, not a complete removal of a whole section.

in bip-0043.mediawiki:None in b75ecca870 outdated

  40 | @@ -41,6 +41,9 @@ from overlapping BIP32 spaces.
  41 |  
  42 |  Example: Scheme described in BIP44 should use 44' (or 0x8000002C) as purpose.
  43 |  
  44 | +Note that m / 0' / * is already taken by BIP32 (default account), which
  45 | +preceeded this BIP.

periodic1236 commented at 7:05 PM on August 23, 2014:

I don't know if you'll keep this wording, but s/preceeded/preceded/.

schildbach commented at 8:20 PM on August 23, 2014:

Thanks, fixed.

prusnak commented at 7:15 PM on August 23, 2014: contributor

The only thing that could take path m/0' is BIP-0000 which does not exist, but I am fine with the note to be included in the spec. ACK.
BIP43: Block usage 0 because it's already used in BIP32. It would be unfortunate if BIP32 wallets would at one day overlap BIP43 wallets. 12b2dc08a4
schildbach force-pushed on Aug 23, 2014
mikehearn commented at 9:11 PM on August 23, 2014: contributor

Well, that's true under the current definition, but BIP43 can be amended to state that zero is special. Patch looks good to me.
laanwj referenced this in commit 75a087e4ae on Oct 15, 2014
laanwj merged this on Oct 15, 2014
laanwj closed this on Oct 15, 2014
ajtowns referenced this in commit 80c6129cee on Oct 18, 2019

Contributors

Linked (view graph)

#61 Remove proposal for wallet structure from BIP32.