User:RobLa-WMF/Conversation slow start

From MediaWiki.org
Jump to navigation Jump to search

The short version: when convincing a group about an idea, start by asking the group one question, then two, then four, until you hit a pace that matches the cognitive bandwidth of the group you are speaking with.

Conversations are like campfires[edit]

Sometimes, starting a having a conversation can be a little bit like having a campfire. Some matters are a bit like lighter fluid; that is, they can be helpful for getting things going, but care must be taken to keep them from becoming all-consuming (Parkinson's law of triviality). Topics can vary similarly to pieces of firewood; some will go quickly, some will need skill to get going and keep going, some will put off noxious fumes.

The campfire metaphor is one way of thinking about it. Another more complicated metaphor is computer network transmission...

TCP congestion collapse[edit]

"Slow start" is something that is generally well-understood by network engineers, for good reason. The Internet as we know it today nearly died because of it in the mid-1980s. From https://en.wikipedia.org/wiki/Network_congestion#Congestive_collapse:

Congestive collapse (or congestion collapse) is the condition in which congestion in a packet-switched computer network prevents or limits useful communication. Congestion collapse generally occurs at "choke points" in the network, where the total incoming traffic to a node exceeds the outgoing bandwidth. Connection points between a local area network and a wide area network are the most likely choke points.

When a network is in such a condition, it has settled (under overload) into a stable state where traffic demand is high but little useful throughput is available, and there are high levels of packet delay and loss (caused by routers discarding packets because their output queues are too full) and general quality of service is extremely poor.

Congestion collapse was identified as a possible problem as far back as 1984, for example in RFC 896, dated January 6, 1984. It was first observed on the early Internet in October 1986, when the NSFnet phase-I backbone dropped three orders of magnitude from its capacity of 32 kbit/s to 40 bit/s, and this continued to occur until end nodes started implementing Van Jacobson's congestion control between 1987 and 1988.

When more packets were sent than could be handled by intermediate routers, the intermediate routers discarded many packets, expecting the end points of the network to retransmit the information. However, early TCP implementations had very bad retransmission behavior. When this packet loss occurred, the end points sent extra packets that repeated the information lost, doubling the data rate sent, exactly the opposite of what should be done during congestion. This pushed the entire network into a 'congestion collapse' where most packets were lost and the resultant throughput was negligible.

The algorithms work by starting "slow", exponentially increasing the speed of transmission based on receiving receipt acknowledgement of the packets. "Slow" is in quotes because exponential growth can be very fast, even from very small starts. Speed increases level off based on acknowledgement of receipt by the receiver. Packet loss is treated as a signal for endpoints to slow down. This innovation allowed early Internet pioneers to stick with the design principle that the network shouldn't meddle with interactions between end points.[1]

Cognitive collapse[edit]

In the same way that networks of computers are vulnerable to congestion collapse, groups of people are similarly vulnerable to cognitive collapse. Information and ideas can flow freely between a well-networked group of people when those ideas are concisely stated. To flog the metaphor, when the information won't fit in a single "packet", the reliability of transmitting that information goes down substantially. Good flow control is needed to transmit the information between people.

Bikeshedding is possibly a symptom of congestion collapse, where someone puts forward something big and important, and people trying to engage start with the tiny piece they can actually understand.

Conversation slow start[edit]

Avoid transmitting information that is too complicated for your audience to understand. Start simple, check for acknowledgement, and then figure out the speed at which your audience is ready to absorb your ideas.

Start by asking the group one question, then two, then four, until you hit a pace that matches the cognitive bandwidth of your information conduit.

RobLa's notes[edit]

Some stuff that I used to understand better that I may incorporate into this later

  • End-to-end principle