Computers are SOO big - V Clunky

[kilemall]

Or that semiconductance differs, so you need to use vacuum tubes.

I so want to impose a TL5 vacuum tube/analog computer as an emergency repair on a low tech world. I do I do I do!

 

[samuelvss]

Or that semiconductance differs, so you need to use vacuum tubes.

Which, IIRC, are resistant to EMP. Not to say that jump induces an EMP, but possibly something analogous. Robots, with their much smaller, sexier brains, need to be stowed in their faraday cages for jump?

IMTU I explain it by, well, not explaining it....

 

[aramis]

Which, IIRC, are resistant to EMP. Not to say that jump induces an EMP, but possibly something analogous. Robots, with their much smaller, sexier brains, need to be stowed in their faraday cages for jump?

IMTU I explain it by, well, not explaining it....

No... IMTU...
Nothing semiconductor works right in jump except electron clouds. And semiconductors that work in jump don't work in N-space. Even neurons don't work quite right, hence jump sickness/madness, anagathic interactions, and the Jagd-Il-Jigd being unable to tolerate jumping.

Mind you, with modern (TL8) micro-vacuum tubes, a modern desktop merely becomes a miniframe. TL10 should be able to get them down to about equal to 1980 semiconductor sizes.

Most of the tonnage, however, isn't the core - it's the redundant systems and control processors.

 

[nobby-w]

Nice bit on the quantum part.

I took this on in my IMTU thread, basically I explained the LBB2 computer pricing and performance as being extremely fault-tolerant and dreamy support levels, and dropping in prices in orders of magnitude but losing licensing for carrying passengers and ultimately freight, increasing instability under duress and bad/non-existent support for swapping in/out starship hardware.

So ya you can run your ship on a PC equivalent but you'll be running under market/illegal rates and find yourself without a working system at the worst possible times.

I've also included sensor systems and antennae in the volume of the computer - it's not just a computer, but an avionics package. Plus, you could make a case for radiation tolerant memory. People used core memory for aerospace decades after it was obsolete for terrestrial applications. It's non-volatile and tolerant to radiation and EMP.

I've seen people comment that folks would fly core in satellites if you could still get it. I've even seen ads in Military Technology for core memory systems, although that was back in the 1980s.

 

[Adam Dray]

Or that semiconductance differs, so you need to use vacuum tubes.

That's just not a universe I want to live in, man!

 

[BillDowns]

I've also included sensor systems and antennae in the volume of the computer - it's not just a computer, but an avionics package. Plus, you could make a case for radiation tolerant memory. People used core memory for aerospace decades after it was obsolete for terrestrial applications. It's non-volatile and tolerant to radiation and EMP.

I've seen people comment that folks would fly core in satellites if you could still get it. I've even seen ads in Military Technology for core memory systems, although that was back in the 1980s.

One big advantage of core memory is that it retains values during power loss. the old AN/UYK-7 used in the original Aegis design did not have to be rebooted if that happened - only the CPU registers had to saved in the core and reloaded from there. Doesn't take very long to restore 24 registers.

 

[kilemall]

One big advantage of core memory is that it retains values during power loss. the old AN/UYK-7 used in the original Aegis design did not have to be rebooted if that happened - only the CPU registers had to saved in the core and reloaded from there. Doesn't take very long to restore 24 registers.

You can get the same effect with CMOS memory.

https://en.wikipedia.org/wiki/Nonvolatile_BIOS_memory

 

[simonh]

Things a 1 space ship navigation computer needs to be able to do all at the same time:

* Plot a course from any point in the system to any other pony in the system. This is reasonably a 'commodore 64' scale problem.

* Calculate the relative positions of billions of registered or detected objects in a system in real time. Now we're a bit outside C64 territory.

* Run simulations on this data set up to many months or years into the future, calculating all possible intersections within programmable tolerances, in a matter of minutes.

* correlate the inputs of all Ensor systems and detect identify and track any sensed objects in real time.

I'd also like to point out that a single example of a 'size 1' program does not constitute the upper limit of what a size 1 program can be. All we know is that it is somewhere in the problem space considered appropriate for a size 1 program. Other size 1 programs might be bigger, others might be smaller. The only way to work out the upper limits to what constitutes a size 1 program is to compare them to size 2 programs.

Simon Hibbs

 

[samuelvss]

No... IMTU...
Nothing semiconductor works right in jump except electron clouds. And semiconductors that work in jump don't work in N-space.

I like it. See I didn't explain it because my basic philosophy is that if they want science they should go play D&D, but also because I didn't want to have to figure it out. But you've the thinking for both of us.

Here's something: I have always had a shipboard preference for lasers as small arms; no recoil for zero-G, and all. But if the electronics work in N-space, but not in in J-space, this would give another reason for using the good ole slugs-n-powder stuff.