Hacker News Reader

Comment by sowbug | Hacker News Reader

Isn't using a company's product exactly when you'd expect a dependency on that company's service?

This is an artificial dependency. There is no good reason why this device should require a specific nameserver. Here are some reasons I can think of that this would be required:

* Getting device metrics for Google's internal use

* DNS lookups that are only available from querying google directly - in other words, recursion is disabled for some records needed by the device

The first case should be entirely optional, the second case is deliberate subversion of internet standards.

I spent some time researching consumer hardware that consistently sacrificed usability for architectural purity. I'll update this reply as soon as I find an example.

I'm trying to figure out how this use of DNS would improve usability for the device owner, and how implementing the DNS client in a standard way could be considered "architectural purity".

You're repeating yourself.

Personally, I expect devices that claim to function with TCP/IP, DNS, https, etc. actually function with them, and not with a tiny subset of their proprietary implementations.

If Google wants to sell captive toys well, that's nothing I'd ever buy, but they're free to. But it needs to be clear that it is a proprietary widget, dependent on Google's services and incapable of operating in environments where those dependencies are unacceptable.

Were you really under any illusion that all set top boxes/dongles weren’t proprietary widgets?

It's not a dependency that I expect of browsing devices. I expect to be able to use eg. a TV, a radio or an ebook reader entirely without relying on the vendor, save for technical support maybe.

Chromecast isn't a browsing device.

You don't have every single car brand with their own fuel at the gas station right?

If a car brand did, I would find it surprising that a buyer of that brand would complain about it.

He would complain if he was forced to buy _only_ that brand's gasoline.

That's in fact the approach of many printer manufacturers regarding toner/ink cartridges, and usually users complain about that.

I wrote "buyer," not "user" or "complainer." I have no problem with self-appointed Good Samaritans who leave anti-Lexmark Amazon reviews. But a buyer of a car requiring the manufacturer's gas would understand what he or she was getting into.

Just a random aside. You couldn’t download third party apps on the 3rd generation AppleTV, but without jailbreaking, the community made a Plex app, merely by running a Python script and redirecting DNS to your computer that intercepted calls to the Apple Trailers app.

https://github.com/iBaa/PlexConnect I’m sure some inventive person could figure out a creative use for a ChromeCast.

But overall, I agree with you.

You're being consistent with Richard Stallman's model of freedom for general-purpose PCs vs. appliances:

As for microwave ovens and other appliances, if updating software is not a normal part of use of the device, then it is not a computer. In that case, I think the user need not take cognizance of whether the device contains a processor and software, or is built some other way. However, if it has an "update firmware" button, that means installing different software is a normal part of use, so it is a computer.

Many of the objections in comments to this post are expressed in terms of standards compliance or interoperability. I think it makes more sense to analyze whether it's a general-purpose PC or not.

https://stallman.org/stallman-computing.html

DNS is a protocol not a service

Respectfully, DNS Spelled out is "Domain Name Service"

There is both a protocol and a service associated with DNS.

I think you meant to reply to the parent comment.

Reply on news.ycombinator.com