Hacker News Reader

Comment by owenwil | Hacker News Reader

This is a pretty smart move, considering that rumors point to Google's "game streaming service" being available on every device that runs Chrome, without additional installation. That would mean a launch across:

* Windows

* Mac

* ChromeOS

* Linux

* Mobile

...overnight. Pretty compelling way to draw in new users given millions of people already own a Switch controller and could start without needing to buy Google's one.

View on HN

Is there a reason to believe this is a Google strategic change instead of some individual Chromium hacker who happens to own a Switch and wants to use its controller for fun? (I don't see any more context on the bug, so it seems to be it could be either.)

The author has a @google.com email address.

Sure, still could be personal interest. I certainly commit things to my employer's source control that are personally interesting and not company strategic priorities (but still correct changes to make) on occasion.

I'm neither confirming nor denying this theory since I have no idea, but googlers can and do work on stuff that has nothing to do with their job function, even aside from formal 20% projects.

Also, I don't know if it's commonplace anymore, but I remember in the past a lot of people that worked on Chrome had @chromium.org addresses, too.

I feel like if it was unrelated to their official work they wouldn't use a google.com address?

That doesn't tell you much - it's not unrelated to their official work, it's just not related to company priorities. If it's a project you work on at / for work, copyright is likely still owned by the employer. See also the number of personal projects that are not official Google projects at https://github.com/google .

No, the policy is to use your google.com address, with some exceptions[1]. This confuses a lot of people, as things done for fun by one person on weekends (maybe not even official 20% time) are mixed in with official projects.

[1]: https://opensource.google.com/docs/patching/

Is this for patches that occur on your own time without company resources? Even in jurisdictions (like CA) where they can't enforce ownership over what you did?

I am neither a Googler nor a laywer, but I think yes and yes - if you don't care about copyright ownership and only licensing (which most of the time is fine - you usually only care if you either want to commercialize it later or you want to make sure either you have standing to sue infringers or Google doesn't), there's an easy process for open-sourcing something and letting Google retain ownership, and a much more involved process for confirming that Google doesn't own it. So it's easier to let Google claim copyright and not argue about whether they really have had ownership.

Some relevant links are https://opensource.google.com/docs/releasing/ for the personal-project-with-Google-copyright flow and https://opensource.google.com/docs/iarc/ for the personal-copyright flow.

No, we use our @google.com address for anything that might fall under the copyright assignment rules of our employment agreement. I won't debate whether those rules are actually enforceable - I have no idea - but if you're actually curious, the documentation is mostly public.

https://opensource.google.com/docs/releasing/

I see, thanks!

> Google's "game streaming service"

A number of companies, both startups and established game companies, have tried this. It's not taken off because you can't beat latency.

Not every gamer cares about latency.

If there's enough of it, yes they do.

Personal opinion: I suspect it's something most gamers care about even if they don't realise they do. If your controls feel disconnected from your gaming experience then you will get frustrated pretty quick.

It seems like the "obvious" way for Google to do this is to have a thin client with smarts instead of just a video terminal, similar to line editing and local echo in terminals (or Mosh's predictive input). The Chromecast is a smart thin client: when you play YouTube videos, it's actually running YouTube in a Chromium instance on the Chromecast, not streaming it from your laptop or phone, even though you control it from your laptop or phone. And with Google's experience in NaCl, WebAssembly, WebGL, etc., it wouldn't surprise me if they figured out how to ship enough of the game locally but still do heavy processing on the server.

Of course, it's technically involved to make a local client for every video game they want to support. With local echo or even YouTube the scope of the problem is smaller.

I used the Google project stream beta. I have a gigabit internet connection but live a few hundred miles from a Google data center. The experience was pretty good. Assassin's Creed is not super latency sensitive, and most of the time I didn't even notice. I'd probably prefer to stream 1080p60 than play at 30fps on a console.

I'm still not convinced it can be commercially successful though: not that many people have a good enough connection, and most people who care about games at all will already have a better way to play them.

The Project Stream test was encoded at 60FPS, but the game was only running at 30FPS.

Very easy to compress a frame that's identical to the previous one. I guess they should wind it up to 120Hz to make it look even better..

Steam in-home streaming over gigabit LAN seems to be the best case scenario for this tech, and for me that is still not good enough for most of the games I like to play.

Nvidia GeForce now has millions of users. What does "not taken off" even mean? I play Fortnite in GeForce now basically all the time.

Reply on news.ycombinator.com