Skip to content
Go to file

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time

qiki -- rate anything

A qiki is part of a no-center public record of what anybody thinks of anything. Including what we think of each other's thinking. And what we think of each other. Crazy, right?

get good

The best opportunity for human betterment today is to get good at deciding:

  1. what is
  2. what ought to be
  3. what to do about it

Let's get better at these. Many great consequences are held back until we do. We don't need to agree. But we do have decisions to make, and move on, in a way most everyone can get behind.

Impossible you think? I'm not so sure. I am sure if we don't figure this out it will ruin us. I think someone's going to figure it out, how to resolve as a group. I'd rather it be us. I'd rather it be humans than some species living on our ruins.

We must find a way to decide what is happening. This is why many discussions talk past each other. We need some agreement on what is happening before we can talk about what matters. We need enough agreement on what matters before we can talk about what to do.

Who is we? This applies to any group you can think of. The opportunity is overripe for getting better at group decisions.


I think the gizmos can help us out. Start by storing our thinking outside the big tech silos. Then anybody can begin looking for patterns. And turning the gizmos loose looking for patterns. And looking for meaning. The time to create value and safety by obscuring will soon be over. It is time to put all our thinking on the table.

A qiki will allow you to rate and relate:

  • anything you can see or say (not just what you're shown or told)
  • to any degree (some things matter more than others)
  • with any verb (not just likes and wows)
  • explanations welcome (the world wants your fresh thinking)

Rate the ratings. Rate the raters. It's everyone's job to have opinions. And everyone's job to figure out what they mean.

No company is powerful enough, no government is rich enough, no review staff is responsible enough -- no subset has enough smarts -- to decide the most important decisions. Or decide who decides them. This has got to be everyone's job. Yes it's going to be very hard.

This is how we'll get big things done without being overwhelmed, without losing our minds, or our humanity (or whatever divinity means to you): by teaching machines to tease out of many opinions what we want, what actually matters to us.

If you don't think machines are up to that you're probably right. We need better machines and better machine teachers. But before we start those negotiations we need a rich collection of opinions out in the open. Let's get started.

((some kind of smooth transition to some technical stuff goes here))


A qiki word, like a natural language word, can represent anything. Unlike a natural language word you can make up a lot of them easily.

When you make up a new qiki word you string together three other qiki words:

  • subject
  • verb
  • object

and you give the new word a number and some text:

  • number (1 means normal, 2 means you mean it twice as much. Numbers are awesome. They can be tiny, they can be huge. And you get to say what they mean)
  • text (because you've got a lot of splainin to do)

and store this in a lex.


A lex is a dictionary for your made-up words. For each word, a lex will remember:

  • id number
  • when you defined it

So that's 7 parts to a word. Here's what they're called:

  • sbj - subject
  • vrb - verb
  • obj - object
  • num - quantify matters
  • txt - name, explain, discuss, UTF-8
  • idn - identified sequentially
  • whn - UTC seconds after 1970

You could call this seven-y thing a sentence. A lex is just a bunch of words. And each word has these seven attributes. (All but txt are qiki numbers.)

hello proverbial world

Let's make up some words. Here is some word-making-upping action in the Python programming language. Before a lex can say hello to the world, it has to define world. And hello. Those are new words. The lex also has a word for itself.

lex = LexInMemory()
hello = lex.verb('hello')
world = lex.noun('world')

Once a lex has a verb for hello and a noun for world, then it can say hello to the world.

lex[lex](hello)[world] = 42,"How are ya!"

The syntax will make a little more sense if you diagram a sentence with subjects and objects in squares, and the verb in a circle.

#  _________       __       ________
# |         |    /    \    |        |
# | subject |-->( verb )-->| object |
# |_________|    \ __ /    |________|

Okay, it doesn't make a whole lot more sense. But you see the resemblance right? Give it time to grow on you.

The hello-world statement is itself a word. Let's look at this new word.

word = lex[lex](hello)[world]

print(int(word.num), word.txt)
# 42 How are ya!

# Word(sbj=lex,vrb=hello,obj=world)

is a word a sentence?

Yes and no. Yes every sentence defines a new word. But no there are some other ways to make up words.

I lied earlier, the subclass LexSentence contains words of seven attributes each, but a Lex is simpler. The only attribute a word really has to have is an idn. A word can be anything you can identify with a number. And a lex can be any kind of access to identified words.

is a lex a bunch of sentences?

Almost. A lex always provides access to a bunch of words. Give a lex an idn and it will give you a word.

idn = 7
word = lex[idn]             # These do
word = lex.read_word(idn)   # the same

If you know a database with numeric ids you can make a lex that will provide a word for any record in that database. Then you can make up sentences that refer to those records. You could triple-like one record, suggest corrections to another, and claim another validates some comment somewhere.

the no-center thing

If you want, your lex can connect to other lexes. Because of dirty tricks with qiki numbers, the sentences in your lex can use the words in any other lex.

If you want help making sense of your lex and your words, the best chance comes from making it all public. Everyone's job, remember. You get to rate how well that works.

Human opinions are messy. That's what we do. You may want to use tools to extract meaning from the mess of human opinions in your lex. Or you may want to make and share tools that do that.

so what does a word represent?

A word can represent a skit about cheese, or that meme your friend would like, or the cure for your grandmother's cancer, or the good part of a video, or whoever designed that curvy part on your phone.

A word can be more than a rating. It can represent any thing, or relationship between things. It can be any abstraction at all.

A word can represent an algorithm you make to extract something interesting from a bunch of other words. Make up a qiki word for that, and other lexes can use it. And rate it. And reward it. And encourage more like it.

A word can represent the next big thing that will make this software obsolete. Something else is bound to eclipse the reckless experiments I try to start with qiki, after they go down in flames from regulation, big ventures, bungled financing, and patent trolls. So please steal this idea and redo it right. Or help me make that more tempting.

who is this for?

I hope you'll use this package to spin up your own lex and integrate it into your site and start collecting opinions and other abstractions. And share them with other trusted lexes. A lot of software needs developing too. This is barely an appetizer of a demo. Just some quarks for the next social physics, as it were.

To start, only geeky developer-y types will be able to create a lex. That should change.


I'm not sure, you can decide, but I think you should be able to change anything except your history of changes. I know, right?

This package needs a lot of change before it can play a part in the ways mentioned here. I hope you'll participate.

-- Bob Stein, Generation 2 is running there now. Generation 3 is what's in this repo.

It's terrible to start over. It's worse not to be able to. Let's start over again.


No description, website, or topics provided.




No releases published


No packages published
You can’t perform that action at this time.