muid 0.2.8

Memorable Unique Identifier

muid

Memorable Unique Identifiers

Wait you say ... that's an oxymoron

Memorable unique identifiers are a provocative misnomer. When generating unique identifiers such as privately used keys, memorability is antithetical to uniqueness. MUIDs might be better termed "hash-memorable" identifiers. They form a subset of UUIDs whose SHA-256 hashes are memorable.

Start mining now

Just want to mine? Run this bash script.

/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/microprediction/muid/master/examples/mine_from_venv.sh)"

The script will explain itself and the motivation, as it mines. If it doesn't work then maybe try

ruby -e "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/install)"
brew install python

although we'd rather not get into a debate about the "best" way to install Python on a mac.

Sponsor a statistician

If you find a spirit animal of length 13 or more you can make the day of someone looking to enter a worldwide statistics contest, to be announced Apr 1st, 2020. The clock is ticking.

Using the library directly

As per https://muid.readthedocs.io/en/latest/ ...

Install

pip install muid

MUIDs are just like UUIDs (uuid.uuid4)

import muid
key  = muid.muid4(min_len=8)  

Only they take much, much longer to generate because ...

... the SHA-256 hash is memorable

Try this:

print( muid.mhash(key) )    

Don't see it yet? Look closer. Here's an example I just happenned to have lying around:

f01dab1eca70403aa0c700f6c29596c4

This reads a bit like:

Foldable Cat  

Mining

It is trivial to mine for MUIDs.

muid.mine()

At time of writing, mining MUIDs is roughly one order of magnitude more profitable than mining bitcoin even if you use this lousy library to do your mining. With a little work, you should be able to mine with 100x the economics of bitcoin ... at least for a while!

Caching in

There is a bid for MUIDs of len at least min_len where:

min_len = int(requests.get("http://www.microprediction.com/config.json").json()['min_len'])

More details on that soon.

An example of the use of MUIDs

We hope you have your own uses and would love to hear about them. Many applications can benefit from one less join.

Implementation decisions

We welcome thoughtful suggestions at https://github.com/microprediction/muid/issues or https://algorithmia.com/algorithms/microprediction/mverify/discussion.

Choice of hash

muid.mhash() 

Uses SHA-256 hash from hashlib.sha256

Readable hex

Hex strings are just a,b,e,d,e,f plus digits and hyphens, which makes for a sucky game of scrabble. We introduce readable hex as follows:

word.replace('0', 'o').replace('1', 'l').replace('2', 'z').replace('3', 'm').replace('4', 'y').replace(
    '5', 's').replace('6', 'h').replace('7', 't').replace('8', 'x').replace('9', 'g')

which is to say

Hex	Human	Hex	HUman	Hex	Human
1	l	5	s	9	g
2	z	6	h	0	o
3	m	7	t
4	y	8	x

with a,b,c,d,e,f unchanged of course.

Collisions (forgettaboutit)

It is well appreciated that approximately 2.71 quintillion uuid4() can be generated before the risk of collision exceeds fifty percent (Wikipedia). Thus UUID collisions are a non-issue. But since an MUID collision requires an underlying collision in UUIDs which are generated while mining, a moment's reflection should convince the reader that the computational capacity required to create MUID collisions over any interval of time is at least as large as the computational capacity employed to create UUID collisions. In short, the thing to worry about is the relatively short supply of MUIDs (an obvious limitation) not collisions between them.

Stupid dog trick

https://vimeo.com/396819347