Hacker News Reader

Comment by lilyball | Hacker News Reader

> Most people familiar with the situation are attributing the problem to a Roku, in that it's the Roku devices that are malfunctioning and not the Nintendo Switch.

I would certainly hope that "entering an endless boot loop" is considered a malfunction no matter what network traffic is occurring.

View on HN

It's a good example of not following Postel's law:

TCP implementations should follow a general principle of robustness: be conservative in what you do, be liberal in what you accept from others.

As originally mentioned here in RFC 761 but credit to Wikipedia for having the citation. https://tools.ietf.org/html/rfc761

> I would certainly hope that "entering an endless boot loop" is considered a malfunction no matter what network traffic is occurring.

Once the switch is in airplane mode, the endlessly looping device seems fixed.

The fact that it reboots is probably a bug.

The fact that it reboots endlessly is caused by the nintendo switch endlessly broadcasting the same signal.

That's still not really Nintendo's fault though. Network services are allowed to repeatedly broadcast the same packets, and it does so for legitimate purposes. If competently written, the Roku should have no trouble discarding invalid input.

>Network services are allowed to repeatedly broadcast the same packets, and it does so for legitimate purposes.

To add to this, even if it wouldn't be for legitimate purposes, a device shouldn't enter a boot loop because another device is malfunctioning.

Malfunctioning or malicious.

Counterexample:

  1. Hello!
  2. Three-way handshake.
  3. Authentication.
  4. Please reboot.
  GOTO 1

If you can auth without intending to that seems like a bug to me.

From GP: > no matter what network traffic is occurring.

The chain described there is "possible network traffic", and would be a valid type of network traffic, "while-meaning-to" if we so please, but there is no way to distinguish that and it is thus meaningless to code.

A device that would never reboot, no matter what network traffic is occurring, would never allow you to reboot it remotely... SSH is network traffic. Your SSH authentication also is.

The point wasn't about the Roku, it was a pointed reply to that particular point, I think. :)

I really wish we could stamp out this excessive pedantry. It is completely clear what OP meant. There is no need to pretend we are robots.

We may not be robots, but our code is still run by machines which are incredibly pedantic. Any imprecissions which require human intuition to untangle won't help solve practical problems.

Hackernews comments don’t get run on anything. Not to mention that they are imprecise anyway because they are written in English.

Pedantry is necessary if we are going to talk about computer bugs and flaws in the intended behaviour of systems (AKA the "hacker" part of "hackernews"). I hope we never change.

I know many hackers who can communicate just fine without redirecting conversation to a debate over the meaning of words. Semantics are trivial to clarify. Pedantry honestly has very little to do with building things except when resolving miscommunication.

Besides, words have multiple meanings, and there are many floating signifiers in the world. Correcting diction without an acknowledgment of intent might as well be pissing into the wind.

On a personal level, aimless pedantry is a terrible attribute in people you work with, and these people can be toxic to productivity.

I really wish people would be more precise in their language. Often, it is completely unclear to me what people mean. I'm too autistic to properly read "beween the lines".

Oh, in all likelihood the actual system doesn't have any auth at all! But there are legitimate reasons for network traffic to toggle the "power state" of a machine (not that I think such features are useful). Some computers implement it. It's just unfortunate they didn't implement it right in the Rokus.

Reply on news.ycombinator.com