Classic Traveller - Computers
TOC
Computers
I’ve been of several minds regarding computers within traveller.
Let me reference the tonnage chart again:
A model 7 computer at TL “D” has roughly ten times the capacity of a TL 5 computer. Given that we’re arguably somewhere around tech level A/B , and how much storage and computational capacity have increased, it’s understandable why the attitude towards this since the early 2000’s has been a degree of “WTF?”.
I’ll preface the rest of this with the following: there is an assumption that “tonnage” here is a catch all for displacement tonnage, aka volume, where it’s also assumed the overall equipment density roughly matches that of 1 ton of liquid hydrogen (14 cubic meters, p.56) or as noted later under drawing deck plans:
p. 67 : One ton of ship displacement equals approximately 14 cubic meters. With a 3.0 meter floor-to-ceiling height, one floor square (1.5 meters by 1.5 meters by 3.0 meters) equals 6.75 cubic meters. Two such squares equal 13.5 cubic meters, or approximately one ton.
So do the equipment sizes make sense?
It’s been an article of faith online since at least 2000 that the tonnage numbers for computers are so much crock. Looking at the average mobile phone these days with enough computing power to embarass a Cray from the 80’s, it’s easy to nod and say yes. I know I did.
Well- TL5 is basically a WW1-2 era mechanical fire control computer for a battleship, or if you handwave a bit and assume transistors can be manufactured with tech a bit earlier than they were invented, the electronic equivalent. This was easily 1 ton. Especially when you talk volume and factor you need room for an operator, stored program cartridges or cards, etc. - punch cards were still common into the 70’s. Even a modern 4-post cabinet with room to rack equipment out, networking hardware, etc., is still easily going to take up at least one displacement ton of space. Actually more if you’re going to get to the sides or back for maintenance
Well, OK, but…..
Space is a brutal and unforgiving environment. All equipment on a spaceship has to deal with:
vibration (general systems, thrust, atmospheric exit and entry, some degree of combat)
temperature extremes - go diving into an atmosphere to scoop fuel. Or see any kind of atmospheric entry per above. And space is cold.
Possible loss of life support - taken apart in orbit for hull repairs, combat damage, dealing with pests including the hijacker variety. This also has attendant issues with loss of temperature cooling / management, exposure to corrosive or exotic atmospheres, and other hazards.
Heat dissipation - vacuum is a wonderful insulator. That heat has to go somewhere (also under reasons why it’s hard to sneak up in space).
Radiation - power plants blowing up, stellar flares, natural background, and cosmic rays.
On top of that, the following is true of most computer systems integrated into ships, as the Traveller rules see it:
Integrated into doors and locks for anti-hijacking protocols.
Integrated into sensors and turrets to enable fire control direction and prediction as well as launching and return fire.
Integrated into power systems for ship control and maneuvering as well as auto-evasion routines.
Integrated into sensors and transmitters for ECM capabilities as well as detection.
As ships get larger - additional terminals to support operation of ships systems from locations other than the bridge (especially power systems) and to facilitate comms with the bridge and between stations.
Integrated into sensors for navigation and jump calculations.
Finally - in the game rules, computing and managing a longer jump requires more computer resources, which in the above volumetric-heavy paradigm translates as much to sensor and integration capacity as anything else - the model/6 computer required to perform a jump-6 has a lot of resources left over even while managing a jump. I could come up with several ways this would be true in case I needed to lampshade it.
In short - it requires a lot of cabling and ducting. More to the point - a good part of the tonnage dedicated to a “larger” computer system is almost entirely the additional sensors, electronics integrations, cooling, and so forth needed for more accurate fire control, evasion, and on and on. Even writing up my post the other day, I was wondering “but come on, even within the last 40 years, processing capability for high-reliability systems has gone up” - and now I look back to my experience with industrial plant control systems and the distributed instrumentation and control nodes for those, and consider the need for airline-levels of reliability for extremely hostile environments, without airline levels of maintenance and constant checks, and I’m less sure.
Yeah, maybe the library program should go away in favor of portable tablets that don’t need spaceship reliability at tech levels above “A”. Maybe at higher TLs the ships historical sensor and drive data is stored in a database that allows what we’d call a modern desktop to poll the instrumentation history for better maintenance diagnostics and trends analysis. But just about everything else on the software list needs deep integrations with ship systems, and thus takes up space, and not just for the raw processing power.
Finally - a dirty secret of the computer world is that, as computing power goes up, so does the bloat in software. This is why later editions of Windows 10 and now 11, and MacOS just seem so much slower on the same hardware. Consider that operating systems and office suites like Microsoft office are exponentially larger than they used to be. Sure, the graphic displays and rendering on your screen looks prettier, but the underlying logic isn’t necessarily getting much more done compared to Word 5.
If anything, the general programming flexibility granted to Model 1 and two computers is generous, given the server rooms I recall from my youth. I remember being excited at the prospect of getting a 20 MB hard drive for a desktop in the 80’s with the idea that I’d never use that much space up - and these days I have Photoshop images larger than that.
The rules on capacity and software in memory
You’ll see that the capacity of a model 1 computer is 2/4, and a 1bis computer is 4/0
That first number is active memory / storage to run live programs, the second is reserve storage - basically the rest of your program library that can be easily swapped to the foreground as the various combat phases occur, without having to wait for the reprogramming phase.
A model 1 computer can move and target (but not also launch) and basically needs to stop maneuvering (or shooting) to calculate a basic jump using generate. A model 1bis computer can simultaneously run maneuver as well as target, and launch (for missiles) while generating a jump plan, but will have to wait to load jump-1 or jump-2 as well as navigation because there is no ready storage.
Relevant rules from p. 70:
Computers are identified by model, and specified in terms of their capacity to process and store programs. Models range from Model/1 to Model/7. Capacity is stated as the size of the CPU (central processingunit) and of available storage for programs. All programs in the computer's CPU are processed simultaneously, while programs in storage are available on a revolving basis to replace those in the CPU as needed. For example, a Model/1 computer has a CPU capacity of two, and an additional storage capacity of four. The computer might have in it six programs (each of size or space 1): return fire, predict-1, gunner interact, auto/evade, maneuver, and target. Of these six, only two (the capacity limit of the CPU) can function at any one time (in one phase). In the laser return fire phase, both target and return fire programs would be required, and only those programs could be used with this capacity CPU to effect laser return fire. During a laser fire phase, as before, only two programs could be used: target is required, but the player could select between predict-1 or gunner interact for the program to be processed, depending on which would provide the greater benefit.
During the computer reprogramming phase, specific programs may be removed from the computer and others inserted. To continue the example above, both jump-1 and navigation would be required for the performance of an interstellar jump. Both programs would be fed into the computer during the reprogramming phase, but only after sufficient space had been cleared (perhaps by removing the maneuver and auto/evade programs).
In short - the capacity isn’t your entire software library. It’s more of a memory/processor threading limit - you have additional storage (whether punch cards, cartridges, or hard drive storage) on which you can keep additional software and load it during the appropriate phase.
This also means that a q-ship merchie with an outsized computer, even leaving aside pilot and gunner competency, can rapidly stack up bonuses with a maximum flexibility of attack and defense options.
Shortly we’ll cover combat, and then maybe try to build a ship.
While you’re at it:
If you like good books, the guys at Pilum Press have a discord server. Drop in, and if you haven’t yet, pick up a copy of everything they have at their website.
If you’re more into games, check out the Arbiter of Worlds channel, and the Autarch Discord server as well. There you will find discussions on ACKs, Ascendant, and a number of other non-Autarch RPGs and games like Traveller and D&D.