17000 years ago, the world's first modern French person stood in a cave and tried to find a way to tell us, three hundred generations into the future, what was on their mind.

The survival of the tribe

the glory of the hunt

the majesty of the animals that shared their world

the stars in the sky

And they drew it, as best as they could. These are the cave paintings at Lascaux in southern France.

The Great Bull is 17 feet long; it's the largest cave art animal anywhere in the world.

The Crossed Bison tells us something about the artist; they understood perspective and the idea of representing three dimensions in two-dimensional painting.

Four thousand years ago, an unknown author first wrote down a story about the king of Uruk, a city in what's now Iraq, and he opened it like this:

I will proclaim to the world the deeds of Gilgamesh. This was the man to whom all things were known; this was the king who knew the countries of the world. He was wise, he saw mysteries and knew secret things, he brought us a tale of the days before the flood. He went on a long journey, was weary, worn-out with labour, returning he rested, he engraved on a stone the whole story.

That's the stone: the first tablet of the epic of Gilgamesh, the world's first surviving work of literature.

Gilgamesh was two-thirds god and one-third man. He fought and defeated the monster Humbaba. He sought eternal life.

He is famous enough to receive the ultimate accolade: four hundred years into our own future, Jean-Luc Picard tells his story to Dathon at Eladril.

The French chap? He liked animals. Don't know a lot else. What's the difference?

Words

(insert a video of President Josiah Bartlet opining on the same theme)

words are good

We talk to computers by pointing and tapping and clicking. Like a caveman. We talk to one another with words; pure, unvarnished text.

There will be some of you now thinking, oh god, what he means by using text is this: the command line.

And you will be thinking, get lost, you horrible Linux person, you'll get my copy of Photoshop when you pry it from my cold, dead hands.

That isn't what I mean. Has anyone here played text adventures? You might be thinking, yeah, I remember them. OPEN DOOR. HIT ORC. Let's imagine something a little more complex.

Well, here's a game. I can RAISE HANDS and it works. AUDIENCE, APPLAUD. (Team sit down.) But what would the source for this look like?

This is the source. It's Inform 7. (Talk through it.) If you're a developer and that doesn't blow your head off, you might want to examine your expectations a bit. This isn't the user experience, it's the developer experience, but developer experiences are UX too.

This is how bots work. There's a computer program, and you send it text messages, and it replies. That's the user experience. And it needs to be lots better.

(Quote from Nielsen Norman Group.) This is where we find out how many people who call themselves UX experts are actually UI experts, I think.

Because the UX, the user experience, of talking to computers with text could be and should be and will be loads, loads better. This isn't about what the rounded corners in your text message app look like. It's about how the conversation works.

How do we help people understand that what they've said wasn't understood? How do we project the idea that you're talking to a bot but that it understands you? We don't want to pretend it's a person, but we do want to suggest you talk to it LIKE a person. Maybe the bot quietly hands off some things to a real person sometimes; how do we make that seamless?

A blinking cursor is not friendly: how do we guide people through a user journey?

This isn't just theoretical. One chap in London set up a chatbot lawyer to talk peolpe through the process of appealing a parking ticket. And with the bot's help, 160,000 parking tickets were overturned. A hundred and sixty THOUSAND. That's what you get if you apply an understandable user experience to something complex like the legal system.

At the moment, bots are like the command line. They understand rigidly formatted and expressed text, and if you get it wrong, you get an error. That's not what language is for.

We don't want to teach people to speak computerese. We want to teach computers to speak people-ese.

Now, there's a whole bunch of complicated algorithms and PhDs available in that field if you're interested in it, but the key point is this: it's not really about the technology. It's about how we present that text. It's about the user experience.

There are a swathe of things out there to help with this. Microsoft spoke at the last Fusion about their Cognitive Services tools. Wit.ai. Google Assistant unifies voice search and Google Now. Framer Design are rethinking the chat UI. Lots going on.

That's what drove Apple's success; there's nothing your iPhone or your Macbook can do that a Windows laptop can't, but everyone can understand how to actually do it. That's why design, why UX, are important.

And it's so, so early for conversational interfaces that you can get into that and make a huge difference. Instead of just reading other people's definition of what good UX is, you can SHAPE that definition. People like communicating, and bots let you communicate to get things done. Not just point and grunt at a button on the screen. Instead of cave paintings, we've got all the nuance and precision and delight of language to play with.

Virginia Woolf said that language is wine upon the lips. Since we're in a bar, wine is also wine upon the lips, so I won't keep you from yours any longer.

I've been Stuart Langridge. Thank you for listening.