It would seem what apps needed the most back then was more RAM and compute and it's still true today, we've just offload ed most of the compute and RAM requirements to GPUs now NVLink/NVswitch/Infiband is what Starlink was back then for $5000 DGX sparks and similar systems only
Starlink from 1984
11–20 of 39 posts
Re: Starlink from 1984
#12It would seem what apps needed the most back then was more RAM and compute and it's still true today, we've just offload ed most of the compute and RAM requirements to GPUs now NVLink/NVswitch/Infiband is what Starlink was back then for $5000 DGX sparks and similar systems only
The vast majority of app's RAM usage requirements kinda stalled in last few years. It's just AI that needs it in gobbles
Re: Starlink from 1984
#13Along these lines, it would be neat to use Raspberry Pi or something to make a "terminal to HDMI adapter", allowing me to have an unlimited number (tens or even hundreds) of external monitors for my laptop, each running a separate terminal-mode program such as a text editor, htop, Bloomberg, etc. I often find fully working LCD monitors discarded by people who have no use for them, and could easily collect five or mor…
If you have enough USB ports, you should be able to get away with using DisplayLink dongles for the video part. DisplayLink doesn't require any additional video processing hardware (unlike DisplayPort) or even USB-C ports, you can run the old DisplayLink standard over plain old USB 2! Not great for playing video, but a terminal app should run just fine.
I think the current driver restricts you to 4 displays but I'm not entirely sure why (you can do 6 on Windows so I think it must be a driver issue).
Of course, with your RPi idea, you could connect one computer to dozens or maybe hundreds of serial ports, but if all you want is five more screens, USB should do fine for your needs.
Re: Starlink from 1984
#14In that era (87/88'ish) I used Quarterdeck Desqview to provide multi-user capabilities on my (then) super-charged 486:
https://en.wikipedia.org/wiki/DESQview
It was possible to wire things up so that multiple users could connect to a separate Desqview instance and have their own local user context, albeit an MSDOS one .. it was quite functional as a multi-user system for a few years until I got on minix-list and heard about that little project Linus was working on, then decided to switch ..
I'm pretty sure there's a modern way to do this with a Linux system packed with video cards and multiple USB keyboard/mice accessories .. one of these days it'd be nice to do that, and get things caught up with how it used to be, SGI-wise ..
Re: Starlink from 1984
#15This is a great reminder that many of today's "breakthroughs" were imagined decades ago. What's changed isn't just the technology, but the economics and engineering needed to make it practical
You didn't read the article. This has nothing to do with satellite-delivered Internet.
True, they probably didn't read the article - but their claim is still relevant.
We really don't have an easy way in the modern era to share our PC's in such a way that multiple users have their own keyboard/mouse/video combination, even though our PC's are more than capable of doing this - and the underlying technology is all right there in the kernel - but distro's just don't seem to be catering to this need, since the market hasn't identified it as something that users want.
But I guess, if there were a distribution which sets things up so that 4 or more people can use the same PC, it'd get wider adoption. I see no reason why this isn't practical, especially in cubicle environments - one fat PC for 4 or 5 people, who are mostly just doing web browsing, document editing, and ssh'ing to things, seems like a very interesting and viable service to deliver, distro-wise.
I wonder which distro's are closer to achieving this target, in the grand scheme of things .. seems to me its only a matter of configuring the X server properly, with multiple independent display units, and getting the USB keyboard/mouse dance started ..
Re: Starlink from 1984
#16Along these lines, it would be neat to use Raspberry Pi or something to make a "terminal to HDMI adapter", allowing me to have an unlimited number (tens or even hundreds) of external monitors for my laptop, each running a separate terminal-mode program such as a text editor, htop, Bloomberg, etc. I often find fully working LCD monitors discarded by people who have no use for them, and could easily collect five or mor…
If you have enough USB ports, you should be able to get away with using DisplayLink dongles for the video part. DisplayLink doesn't require any additional video processing hardware (unlike DisplayPort) or even USB-C ports, you can run the old DisplayLink standard over plain old USB 2! Not great for playing video, but a terminal app should run just fine. I think the current driver restricts you to 4 displays but I'm n…
DisplayLink creates virtual displays on CPU and stream compressed feed over USB. That takes non-negligible amount of CPU time and RAM bandwidth, and it makes sense to give it a max concurrent stream count(set aside which of 4 or 6 are more reasonable).
Re: Starlink from 1984
#17StarLAN was a 1Mb twisted pair networking system that was pre twisted pair ethernet
Re: Starlink from 1984
#18This is a great reminder that many of today's "breakthroughs" were imagined decades ago. What's changed isn't just the technology, but the economics and engineering needed to make it practical
I’ll need a moment to let this sink in.
It’s not just technology, it’s a movement.
Re: Starlink from 1984
#19I'm guessing it's called StarLink because the design uses a Star network topology: https://en.wikipedia.org/wiki/Star_network
This is a great demonstration of why trademarks aren't forever or universal, and shouldn't be.
Re: Starlink from 1984
#20Earlier quoted context omitted.
You didn't read the article. This has nothing to do with satellite-delivered Internet.
True, they probably didn't read the article - but their claim is still relevant. We really don't have an easy way in the modern era to share our PC's in such a way that multiple users have their own keyboard/mouse/video combination, even though our PC's are more than capable of doing this - and the underlying technology is all right there in the kernel - but distro's just don't seem to be catering to this need, since…
We really abandoned the idea when computers got so cheap that you could just have a separate computer at each station. Mini PCs are $200, how will your solution compete with that? Thin clients won't get much cheaper than that as they have all the same hardware. Thin-client-less multi-user systems might, but the extra administration complexity ($howmany per sysadmin hour?) will make it a tight squeeze and you may as well just stick with what's familiar.
You could have more luck with AIO PCs.