Do we like this way of representing the NSEC to facilitate handwritten backup and typing back into the app?
```
NSEC1-AGZ4-Q7LG-WN8V-4K5M
FA4KT-YK8G-VY0R-AHV2-CPL2
KJQ6M-SP8X-PZV2-PQX7-8ZKF
```
Discussion
I like it! Easier on the eyes to have it displayed that way
Yes.
Why not those BIP recovery words? Is that intentional? If it's about writing down, Crockford's base58 is prly more compact and drops just a few chars that look too much like other chars, like i, I, l, 1
NIP-06 even defines this, including a BIP-0044 path and SLIP-0044.
There is no way to make seed words from existing nsecs using NIP06. That's why no one uses NIP06. You would have to start your profile from zero again.
You can decode bech32 to the raw bytes, and the raw bytes into the mnemonic, right? The raw bytes are the secret key, so from there create a mnemonic? Am I missing a step in the process?
You can represent them as seed words, but that's neither BIP39 nor NIP06.
The other issue is that if we have two seed word methods, people wouldn't know which one is which.
Okay, I'll look at the specs to see what the missing step is then. That's fine.
From binary representation to base58 should give a significant shortening of the key. But I'm not sure you're really looking for an alternative representation. The pairing you showed is okay to make it easy to copy by pencil. I'm guessing that's what you're aiming for.
yea but if your ASCII matrix isn't crossreferenced with your retro encabulator the key won't conjoin with the panametric fan. clearly risking a 256 bit singularity. I wish you idiots would think before you post.
It's not as complicated as it seems. Sometimes raw bytes are transformed into alphabetic characters to make them easy to communicate, possibly with a small checking value included to prevent typos. But some texts are more convenient to work with on screen, and others to write down.
Then there are the many flavors of transformation.
I follow you. And I am tangentially aware of what your talking about. My joke was more about the ease and expertise with which you are both conversing about it. And how it sounds to an outsider. a gift I have that many lack is knowing what I don't know but grasping the fringes of a topic. The gift I often lack is the perseverance to learn more haha. Nothing a little adderall won't fix.
This is how I feel trying to follow this
We could also use NIP-06.
Much better!
Looks great!
I wonder if there is a way to convert the nsec to a Bitcoin seed and back?
You cannot convert it into a Bitcoin *seed*. You can, however, simply convert the npub into a p2tr address and vice versa, since they are both created the very same way.
You can derive bip32 seed words and use these to recover, but I think it would be more fun to use hierarchical deterministic derivations instead of just a single nsec
Reminds me of product CD keys 😂
I don't hate it, but seed words is probably more familiar to the modern age.
You can get 75 hours of AOL FREE!!!
There is no way to make seed words from existing nsecs using NIP06. That's why no one uses NIP06. Users would have to start your profile from zero again.
I assume I'll eventually have to migrate to another scheme. maybe with frostr or ephemeral identities. We should support old nsecs but try to nudge toward more future-proof backups.
So my current nsec can get a seed backup, but instead of words it'll be the above format?
Or do we still need to create a new nsec?
Yep, the idea for this format is to use with your current nsec.
Afaik you can use your nsec as entropy and you will obtain a 24 word seed phrase right?
the project of nostr:npub1fjqqy4a93z5zsjwsfxqhc2764kvykfdyttvldkkkdera8dr78vhsmmleku https://github.com/mleku/wordstr
this is the reference implementation of my proposal for a 256 bit exact word key encoding with 7 bits of check and 24 words
i have backed up my nsec on paper using it... you can at least use this tool to back it up even if the guardians of the nips don't approve the encoding
the algorithm is very simple... it turns your nsec into a giant 256 bit number, and then divides it over and over again by 2048, and the values that come out refer to a word in the BIP-39 word list and that's it... there is one extra 8 bits at the front, and i made it so the first (biggest, on the left) bit is always 1 so no matter what the first one returns something other than 0 because 0 is "abandon" and i don't like that word
yes you could use your nsec as entropy but it would be a seed and only like 192 of the bits are used
nostr:nprofile1qqsyvrp9u6p0mfur9dfdru3d853tx9mdjuhkphxuxgfwmryja7zsvhqpzamhxue69uhhv6t5daezumn0wd68yvfwvdhk6tcpz9mhxue69uhkummnw3ezuamfdejj7qgwwaehxw309ahx7uewd3hkctcscpyug I have tried a library that convert any string instead of a mnemonic words to create a new xprv, could it help?
Listen 🎧 here friend
I have no idea what you’re talking about
We the same age … ish
But if we’re the same age
Definitely some others ain’t as hip to this Jive speaking
Seed words
And obviously as we rescue others
Raised without the Internet, etc. this is something that’s gonna have to be taught
🤣🤣🤣
Like 👍
The dash may make users think it needs to be included
I like it, but I would prefer to see more clients adopt NIP-06 seed words.
Only if the UI includes the dashes already
It’s the “ones, L’s, I’d,zeros and O’s that get me.
1lI0O
(One,little L, capital i, zero, ohh!
i do
Yes
"Please enter your windows key to continue..."
No, not so much. Bip39 is better than this
Why we don't just use mnemonics, and generate many nsecs from it? https://github.com/jaonoctus/nip06
We have it implemented, but no one uses it.
Current users can't generate seed words from the existing nsecs. They have to restart from scratch.
It seems that NIP-06 doesn't support multiple keys in the sense that a new key is used for each new note/event. But different account numbers are a thing.
12 words seedphrase is the way.
Amber already implemented it btw. 
Only works for new accounts. You need to start from scratch if you want to use seeds as predicted by NIP06
Why is this?
Nip06 is based on the Bitcoin seeds. The main point of bitcoins version is to make it impossible to figure out the seeds from any of the private keys related to a UTXO.
But we could still just create bip32 seed words from a single private key to use for a backup/restore, just can't derive the existing key as part of an HD set?
Is there ambiguity w.r.t 1 vs I and 0 vs O.
Ya
If we can’t have seed words then yes, this is far better than what we have now!
(Reminds me of MUUN wallet backup)
If its 10 groups of letters/numbers that have to go in order then not having one single string may be confusing.
15*
ALL CAPS NSEC
As one who can type on a computer
Keyboard’s are different than phones
Number & letters may be
WAIT
CAN we color code the numbers to make them stand out
Signal matters