Hacker News new | ask | show | jobs
by wizofaus 1405 days ago
Actually sounds pretty good, but how feasible is it for a phone sized device to be sufficiently high spec to run, e.g., fully featured development tools etc. I mean you need 32Gb RAM, 8 core processors and 512Gb SSDs these days just to run the basics (maybe less if using linux)...if you're doing modern game development or CAD stuff probably twice that. But... something you could dock your phone into as needed to allow a seamless experience that supplied the extra resources as needed might be feasible?
2 comments

I think if/when we get to the point where people with browser based workloads can just sit down at a desk and have the monitors and keyboard automatically connect to their phone, more hardware intensive software will quickly move to the cloud.

You’ll be doing game development and CAD stuff with your phone acting as a thin client.

Yeah, I've tried doing cloud- based development. It's a no from me. I'd think that's further off being universally workable than the phone-docking idea. Though if you've already decided to get fully in bed with a particular cloud provider and totally tie yourself to their proprietary protocols etc. I can see the advantage.
Yeah I think that's a matter of software, plus optimization for a particular use case. Looking at Geforce Now/Stadia/Xcloud/Luna, the underlying technology exists, it's just a (not so small) matter of updating RDP's video stack to be seamless.
It used to be the only way to do UNIX development, many moons ago, via thin terminals.
And you'd voluntarily go back to that?
Yes, that is what cloud computing is all about, with the added benefit making UNIX irrelevant.

Now it is all managed runtimes, on top of type 1 hypervisors, and containers, with browser based shells and IDEs.

Unikernels have won even if it doesn't look like it from the surface.

I don't have an issue with unix but I still rely on being able to work in scenarios where I don't have a reliable/fast-enough internet connection to be able to run multiple hi-res GUI desktops remotely. If we get to the point super-low-latency network connectivity is as available as electricity, then sure, it's feasible. But even 30ms latency can make GUIs feel clunky to work with. I know some devs seem to tolerate it though.
How many people who own a smartphone need to use "fully featured development tools"?