Topic: easyseed(1) secure, multilanguage tool for BIP 39 mnemonic & seed, BIP 32 xprv (Read 459 times)

<1> It remeber me vaguely on some debate while the implementation of UTF-8 ( Ken Thompson and Rob Pike ) maybe there are some valuable lessons in computing history to remember...

<2> Regards to Bech32 I am trying to compile it on Visual Studio 2017 community edition

Why, what’s the real-world use case of BIP 39<1>?  To ease human memorization, transcription, and communication of pseudorandom strings!

Bitcoin has been a leading innovator in the development of better ways for humans to interact with cryptographic gibberish.  With BIP 39 mnemonics, a mere mortal human can memorize a pseudorandom string, write it into a sealed Last Will and Testament in a manner easy to transcribe back into a computer—or if necessary, whisper it into somebody’s ear...

I should think that people may want to memorize, transcribe, speak, or whisper .onion addresses, too.

As an English speaker, which would you prefer to deal with in your capacity as not being carved out of silicon:  “wlupld3ptjvsgwqw” (ouch!), or “real element glow tennis pluck museum hair shuffle”?  Chinese people may prefer “潔 愛 唱 仰 淚 吳 乎 怒”.

For overlapping reasons, I have also been working on the application of BIP 173 Bech32 encoding for .onion address data (more).  That has the advantage of error correction, and a more compact format; different tools apply to distinct use cases.  Again, Bitcoin technical innovation opens new possibilities in other applications of cryptography.

I am also using these methods with PGP.  I’ve seen PGP mnemonics before, of course—but the mnemonic standard I recollect having seen many years ago used a 256-word (8-bit) alphabet<2>, rather than the 2048-word (11-bit) alphabets from BIP 39.  Thus, unnecessarily long word strings.  I am also working on a spec for what I call Bech32 “PGP Descriptors” to encode key fingerprint plus metadata.  See those ugly hexadecimal strings in my signature?  Bitcoin will provide a better way.

Meanwhile, for use in Bitcoin, I have almost completed full BIP 39 implementation in easyseed(1).  It would have been done yesterday; but for the missing final piece, generation of the output seed, I need means to perform the specified Unicode NFKD normalization without dragging in ugly dependencies.  The rest is trivial.

What's the purpose of this feature? It seems (superficially) that there's no real-world use case.

To celebrate the New Year, I have the gift of a hidden unstable, unsupported feature for you .onion-lovers.  Here is the address of Wikileaks (http://wlupld3ptjvsgwqw.onion/) encoded as an 8-word mnemonic phrase in 8 languages or writing systems:

Sorry about that. Sometimes our spam-detecting systems miss the mark and you were accidentally flagged in the process. I've gone ahead and removed the flag and you shouldn't see that message again.

Subject: 404 on my profile and projects!

In a non-logged-in browser, I am receiving a 404 “Page not found” error for my own profile and projects!  In a logged-in browser, I can see my profile and projects; but a project with 10 total open/closed issues shows 0 issues, total.


- 404: https://github.com/nym-zone

- 404: My projects pages, under that URL

I checked status.github.com “All systems reporting at 100%”) and the @githubstatus Twitter feed (last update 2017-11-27, “Everything operating normally.”).

Please fix this.  Thanks.

[email protected]

I like this idea you have.

Personally I am a believer in the entropic properties of dice.

Also your comments lead me to point out a serious flaw in the use of off line web pages.

Namely, shouldn't such a web page (a computer program) not operate unless it was off line?

I have released an initial version of the easyseed(1) utility for secure generation of BIP 39 mnemonic seed phrases.  As any worthwhile software, it comes replete with a manpage, q.v.

My original motivation for writing this was that I needed a lightweight, reliable BIP 39 seed phrase generator with easily auditable sources and minimal dependencies for use on a stripped-down airgap machine.  The source code is short, easy to read, and lovingly commented; it can be readily understood by anybody with a basic knowledge of the C programming language.  Its only dependencies are cc(1), make(1), and a library SHA256 implementation—available on most platforms via libcrypto or otherwise.

Now that it’s written, easyseed(1) is also the first necessary component for my campaign to urge that users stop using saved webpages to generate their Bitcoin magic bits.  What kind of an airgap machine has a web browser installed, anyway?  But most importantly, as a rule of thumb, Javascript code cannot reliably acquire proper entropy for generating random numbers.  This is a persistent general problem, and specifically subject to extended fretting by the author[1] of the most popular BIP39 webpage.

easyseed(1) reads bits straight off /dev/urandom, or from user-provided keymat.  Gathering and processing of entropy is properly the kernel’s job.  My userland utility will let the kernel do its job.  Since it’s written in C, easyseed(1) can reliably obtain kernel-provided randomness on every Unix/Linux platform in about two lines of code (open(2), read(2), plus error checks)—rather than cooking up some tortuous “random” scheme which may or may not perhaps probably work sort-of.

This is a beta-quality initial release.  It is not yet feature-complete:  In particular, I have code partly written to add support for all languages which have wordlists in the Bitcoin BIP repository (currently Chinese (Simplified and Traditional), French, Italian, Japanese, Korean, and Spanish, in addition to the current English).  easyseed(1) does have basic test vectors copied from the Trezor repository; and with those vectors, it runs a self-test on every usage.  However, more tests are desired.

Licensing includes a Bitcoin Consensus Clause, to prevent use by scamcoin pretenders.

I am here opening a Bitcoin Forum thread for discussion of this utility; over time, I will edit and update this post as appropriate.

---

1. Though that is not nearly in the same league as boneheaded absurdity from ignorant developers who confuse multiple distinct meanings of the word “entropy”.  pointbiz: “Perhaps more entropy can be gathered using techniques used on Panopticlick”.  #facepalm  cantonbecker: “I like this idea”.  pointbiz: “I used all the easy techniques from Panopticlick to gather entropy. [...] I added up the low and high entropy bits and my personal results are 34.3 to 42.8 bits of entropy.”  Oh dear heavens, are you using this to generate keymat for Bitcoin!?  Some people should be enjoined with a permanent restraining order forbidding that they ever approach within one hundred metres of crypto-related code.