📅 Original date posted:2021-07-03 📝 Original message:On 7/2/21 11:24 PM, David ...

Andrew Chow [ARCHIVE] /

npub1fgn…ak44

2023-06-07 22:56:00

in reply to nevent1q…cnyu

📅 Original date posted:2021-07-03
📝 Original message:On 7/2/21 11:24 PM, David A. Harding wrote:
> Is there any chance we can take this opportunity to make "h"/"H" the
> preferred aliases? Using "'" in bourne-style shells is very
> annoying[1], and I suspect it's also creating unnecessary complications
> elsewhere.
I've updated the text to use "h".
> Alternatives:
>
> - Completely kill "'" (I'd prefer this, but I realize it's complicated
> with descriptors already being used widely). If "h"/"H" are made the
> preferred aliases, maybe it'd be enough to make implementing "'" a
> SHOULD rather than a MUST; this would push implementations towards
> displaying descriptors using the h versions for maximum compatibility.
Since there already are software implementing descriptors, I don't think
we can do this. I'm not sure about making "'" a SHOULD either.
> - Calculate the checksum over s/(h|H)/'/ (again, I know that's
> complicated with descriptors already widely used)
This has been discussed in the past and the conclusion was that the
checksum should be strictly over the string itself. This would allow for
dumb checksum checkers which don't have to be able to parse descriptors
in order to check the checksum.

Thanks,
Andrew

>
> Thanks,
>
> -Dave
>
> [1] https://github.com/bitcoin/bitcoin/issues/15740#issuecomment-695815432

Author Public Key

npub1fgnnmg7f4wzup9hct8nv5pnd9l07wcjqdjku9ax432n4g69v4rgq7xak44

Show more details

Published at

2023-06-07 22:56:00

Kind type

1 Short Text Note

Event JSON

{ "id": "9d7c32c5ba4ae4faf64a5e0d0e56815e3e10bb0ffabbed574acb76ac0f7a0072", "pubkey": "4a273da3c9ab85c096f859e6ca066d2fdfe762406cadc2f4d58aa75468aca8d0", "created_at": 1686178560, "kind": 1, "tags": [ [ "e", "ccc742843124363b85277237d89f5eb620638480799eb332b8eef2b5e2e21b23", "", "root" ], [ "e", "7135cfe5f58975f5c61ca80447682888c0ea2ef77f3163f97ac2bfb9bba3b385", "", "reply" ], [ "p", "d3574a24208f4e3d0821bb4a69a0c3ae842043d444fa5c4a8c49c369918a6fb2" ] ], "content": "📅 Original date posted:2021-07-03\n📝 Original message:On 7/2/21 11:24 PM, David A. Harding wrote:\n\u003e Is there any chance we can take this opportunity to make \"h\"/\"H\" the\n\u003e preferred aliases? Using \"'\" in bourne-style shells is very\n\u003e annoying[1], and I suspect it's also creating unnecessary complications\n\u003e elsewhere.\nI've updated the text to use \"h\".\n\u003e Alternatives:\n\u003e\n\u003e - Completely kill \"'\" (I'd prefer this, but I realize it's complicated\n\u003e with descriptors already being used widely). If \"h\"/\"H\" are made the\n\u003e preferred aliases, maybe it'd be enough to make implementing \"'\" a\n\u003e SHOULD rather than a MUST; this would push implementations towards\n\u003e displaying descriptors using the h versions for maximum compatibility.\nSince there already are software implementing descriptors, I don't think\nwe can do this. I'm not sure about making \"'\" a SHOULD either.\n\u003e - Calculate the checksum over s/(h|H)/'/ (again, I know that's\n\u003e complicated with descriptors already widely used)\nThis has been discussed in the past and the conclusion was that the\nchecksum should be strictly over the string itself. This would allow for\ndumb checksum checkers which don't have to be able to parse descriptors\nin order to check the checksum.\n\nThanks,\nAndrew\n\n\u003e\n\u003e Thanks,\n\u003e\n\u003e -Dave\n\u003e\n\u003e [1] https://github.com/bitcoin/bitcoin/issues/15740#issuecomment-695815432", "sig": "80262d4791a589380a4315af498b2867de29d5fa3de5f865de749be42e2d7486475086d55ba752a8b3400e3168a26eba7577f4114da02909836461b14ea39796" }