[ih] booting linux on a 4004

[Lixia Zhang]

> On Oct 2, 2024, at 9:25 AM, Michael Greenwald <michaelgreenwald58 at gmail.com> wrote:
> Hi Lixia! (Yes, it's been a long time. At least a decade or 2)
> My memory is (no surprise) cloudy. For some reason I thought you arrived a year or two later than '81.
> I have no memory (again, no surprise) of what, specifically, your promising "alternative" was, but an early reference to (virtual)clock based control of packets sounds plausible.
> 
Thanks for remembering the Virtual Clock work! 
That's actually about packet queuing order at routers.

The host packet pacing idea was sender control. I believe that Van said rate-control wouldn't work but window-control would was because window control has built-in feedback loop (the ACKs); he might have thought rate-control as open-loop (which was not the case, though no one figured out how to use source quench feedback).  
> I'm not shocked by the difference between Van's comment back then and BBR now. There's never a guarantee that claims that something "simply cannot work" or "will obviously help" back then (or even now) will hold up over long time periods, unless there is a formal proof or a lot of experimental validation. (And even then conditions change).
> 
> About the work that I thought predated you: Dave Clark remembers that there was an undergraduate who did work on Source Quench, but not his name, nor details of what he did. So at least I didn't completely hallucinate....
> 
I had some vague memory about this too, and the problem was unsure the right thing to do when receiving source quenches.  

Lixia