Comms Theory for Humans - part 2

Comms Theory for Humans - part 2

(In which I once again get excitable about how communication theory can tell us a lot about how to human)

I talked last week about the maximum capacity of a communications channel. In practice, no channel is perfect, and there are lots of different engineering approaches to working round the imperfections. Going digital made a huge difference, because it enabled information to be sent in little packets, most of which would probably make it through. Rather than send the encyclopedia as a single book, post each entry as a separate letter, type of thing. Less efficient, obviously, but easier for the postie and, this is the important bit, you shift from all-or-nothing delivery to most-stuff-delivered. If things go missing and that bothers you, you can write back and say 'Please resend everything between Mongolia and Nematode'.

Why do humans care? Because ... we do the same sort of thing. Let's have a look at two of the most popular protocols - TCP and UDP. They both use IP, the internet protocol, to get their packets to the right address, but they handle the problem of missing packets in very different ways.

No alt text provided for this image

TCP is a very cautious protocol. There's a handshake (SYN, SYN ACK, ACK, or synchronise request, acknowledge sync request, acknowledge the acknowledgement) which in people often looks like this.

'Hi, how are you doing, fellow human? Got a minute?'

'Sure, let's sit down'

'Thanks, here looks good'

and then formal checking within the protocol that the numbered packets arrived in the right order, with resend requests if any went missing. In networks it's used for information which isn't time critical, but needs to be right - web pages, emails and so on.

In real life, I use TCP a lot with my children.

'Are you okay with that?', 'Tell me that again?', 'Tell me what I just said.'

I also use it in important meetings.

'Can you play that back for me so I can check we're both on the same page?'

UDP is much more reckless. A request comes in, response packets are flung into the ether and any that get lost are quietly forgotten about. Speed is more important than accuracy.

In networks, it's used for time sensitive information. Voice packets are a classic example. It's more important you get most of the data than that you get every last bit, because you want the call to flow smoothly.

In real life, UDP is how we do informal chats. There's still a handshake, but it's just an invite to send data, not an agreement to track every package. Quite often, the data we're transmitting isn't as important at the fact that we're sharing a connection with someone. And that's fine - some of my favourite interactions with people are just sharing.

'Hi, how are you doing?'

'Oh man, it's been a week! The dog needed his jabs, we were supposed to be having friends round for a barbeque but then it rained so much? Did you get much rain? Oh, and we managed to get the holiday booked in the end. On the plus side, my boss has been out this week so it's been quiet at work. And I heard from Emma about …'

So finally.

When you're talking to someone, and it's not working, think about what protocol you're using. Have you hand-shaken that you're both using the same one? If you've got something important to land, are you requesting ACKs? Are you sharing data, or sharing a connection??It could make all the difference.

要查看或添加评论,请登录

Cal Brown的更多文章

  • What would Marcus Aurelius do?

    What would Marcus Aurelius do?

    As you glaze over for the fourth time in fifteen minutes, and start reviewing emails to keep an alert face for the…

    2 条评论
  • Where do decisions get made?

    Where do decisions get made?

    Are we rational beings? Or are we post-rational? Case studies on people with memory defects have shown that we remember…

    2 条评论
  • Network Resilience for Humans

    Network Resilience for Humans

    How resilient should you be? It's not as odd a question as you might think. Resilience costs, but lack of resilience…

    4 条评论
  • Troubleshooting for humans

    Troubleshooting for humans

    We're all just pot plants with complicated emotions. I was reminded of this at the weekend, and I got to test something…

    4 条评论
  • Comms theory for humans, part 1

    Comms theory for humans, part 1

    (Bear with me, I'm going to get excitable about one of my favourite equations here and how it's as valid for humans as…

    3 条评论
  • Underwater cable laying - how hard can it be?

    Underwater cable laying - how hard can it be?

    One of the things I love about managing this unit is the massive range of challenges we face across the teams. This one…

    3 条评论
  • Ask or Guess?

    Ask or Guess?

    There's a theory I'm fond of, that there are two distinct cultures in the world. One culture is very clear and explicit…

  • Boundary Setting

    Boundary Setting

    Books on my reading pile at the moment include The Goal (Eliyahu M. Goldratt), and The Phoenix Project (Gene Kim).

    4 条评论

社区洞察

其他会员也浏览了