Computers

[coliver988]

and I am always reminded how large the computers are in Star Trek. On the Enterprise D they take up several decks, and in fact, there is a duplicate on the other side to counter-balance the weight. That big purple column is the computer. For me, it is just part of the Traveller Universe and part of the game. Later editions (Mongoose, T5) drastically change the computer paradigm to better match real-world changes. But I like the pulpy origins and ideas and feel no need to justify or change things.

 

[Condottiere]

Seems to be mostly used as weapon system control and predictor.

For jumping, you can really specialize on that specific function, at a lower tier.

 

[Rupert]

My take is that computers for spaceships are massive because they must be extremely reliable, so they have lots of redundancy and don't use the smallest (and most radiation sensitive) micro circuitry. They contain their own 'life support' (aircon, etc.), and that adds volume and mass to even a small system. They include interface systems, both terminals to interface with the crew (and those aren't going to get smaller with improving tech) and with the ship itself - and there's a limit to how much a system that's controlling megawatts of power can be miniaturised, even when power losses (and thus heat production) is cut to almost nothing.

Basically they are mission critical industrial control systems, and are designed and built accordingly.

 

[Badenov]

If you are specifically looking at CT / HG, one of the things that has been suggested before is that the "Computer" can be considered to include ALL of the Electronics and their associated hardware (including sensors, communications, countermeasures, antenna arrays, etc) since these are not specified in design sequences in either Book 2 or HG.

This notion of 'the entirety of your electronics configuration including sensors and comm gear and navigation aids' is supported by the 'Relative computer size' modifier in combat. Jump Programs are some of the least expensive programs, which implies they're among the simplest and/or most straightforward. If you know the position of things in the galaxy, relativity's just a math problem. But tracking dozens of maneuvering opponents in a furball in real time and giving a firing solution that's accurate at any sort of range is complicated, which is why Predict 1 is almost 3 times more expensive than Jump 6. But knowing where targets are, and where they're going, with great precision, is absolutely an edge in a fight, and so each new TL brings additional tools to give you a +1 vs the previous TL. Computers are massive as TLs increase because they're not just one computer. It's a suite of integrated sensors and processors.

 

[mike wightman]

This notion of 'the entirety of your electronics configuration including sensors and comm gear and navigation aids' is supported by the 'Relative computer size' modifier in combat. Jump Programs are some of the least expensive programs, which implies they're among the simplest and/or most straightforward.

Just to play devil's advocate for a bit, I think that has more to do with the authors wanting to make the combat programs more expensive rather than complexity. The skills required and the throw target number is a better measure of complexity - along with program size.

If you know the position of things in the galaxy, relativity's just a math problem.

It's a bit more than that. We can'r "solve" the n-body problem without cheating, a jump space plot has to be a general relativity based hyperdimensional n-body problem. How that is easier than a targeting program... I don't see how, since the targetting program is a simple relative movement problem.

But tracking dozens of maneuvering opponents in a furball in real time and giving a firing solution that's accurate at any sort of range is complicated, which is why Predict 1 is almost 3 times more expensive than Jump 6.

I really don't see how this is more difficult. Our fighter computers can do that sort of thing in real time while running the AI piloting the F-16, while solving a hyper-dimensional general relativity based n-body problem takes a lot longer.

But knowing where targets are, and where they're going, with great precision, is absolutely an edge in a fight, and so each new TL brings additional tools to give you a +1 vs the previous TL. Computers are massive as TLs increase because they're not just one computer. It's a suite of integrated sensors and processors.

Agreed.
It's a shame the LBB:2 computer programs don't have a TL component, maybe better programs are available at a higher TL or are smaller at a higher TL...

 

[Badenov]

Just to play devil's advocate for a bit, I think that has more to do with the authors wanting to make the combat programs more expensive rather than complexity. The skills required and the throw target number is a better measure of complexity - along with program size.

To be honest, it's probably more like 'Star systems aren't trying to evade' that makes Jump cheaprer than Predict.

Agreed.
It's a shame the LBB:2 computer programs don't have a TL component, maybe better programs are available at a higher TL or are smaller at a higher TL...

That is indeed a thing that is missing, although it is possible to imagine that there is one sort of Gunnery program, and the 'levels' are just adapting it to use the additional information available at higher TLs. But that doesn't quite square with the RAW

 

[mike wightman]

To be honest, it's probably more like 'Star systems aren't trying to evade' that makes Jump cheaprer than Predict.

No, but you have to solve hyperdimensional general relativistic n-body problems to plot the jump. The system you are in is an n-body problem, the system you are jumping to is an n-body problem (based on a model of that system since the data for the position and velocity of every object in that system is several years out of date and has to be modeled - ask your AI of choice how far into the future our best models of our solar system are good for ). Frame dragging, gravitational perturbations, spacetime curvature in both locations, occluding objects.

That is indeed a thing that is missing, although it is possible to imagine that there is one sort of Gunnery program, and the 'levels' are just adapting it to use the additional information available at higher TLs. But that doesn't quite square with the RAW

Where there is a gap in the rules there are always referees willing to house rule

 

[Badenov]

No, but you have to solve hyperdimensional general relativistic n-body problems to plot the jump. The system you are in is an n-body problem, the system you are jumping to is an n-body problem (based on a model of that system since the data for the position and velocity of every object in that system is several years out of date and has to be modeled - ask your AI of choice how far into the future our best models of our solar system are good for ). Frame dragging, gravitational perturbations, spacetime curvature in both locations, occluding objects.

Well, yes. We're also at TL7. Or maybe 8. Hyperdimensional general relativistic n-body problems are still just math problems, if more complex than what we do in RL today. Though there is one interesting assertion made in HG that is not stated explicitly: A Model 1 computer is TL 5, and at 1 ton, might be expected to be the equivalent of something like ENIAC but is actually much smaller than ENIAC, and it can still run Jump 1 software. Presumably that means Jump 1 software is not as complicated as we might expect. Though it can't get you closer than 100 diameters, so presumably that's a limit of the calculation.

Where there is a gap in the rules there are always referees willing to house rule

Well, HG overwrites the whole software issue with +/- Computer Rating and I guess presumes everyone's getting the best software possible for their computer.

 

[Condottiere]

More likely, you're programming the jump drive.

Synchronizing the dimensional vibrations.

 

[mike wightman]

Have you ever tried to solve the equation for a simple harmonic oscillator in n-dimensions coupling with other particles?

 

[Rupert]

That jump programs are (relatively) cheap, but large (especially for high jump numbers) suggests to me that they're fairly simple, but doing a lot of brute-force calculating. Hopefully this is because they're tracking and allowing for many moving parts, and not because they are Excel sheets using vBasic macros written by over-worked, under-paid, self-taught office clerks.

 

[Badenov]

That jump programs are (relatively) cheap, but large (especially for high jump numbers) suggests to me that they're fairly simple, but doing a lot of brute-force calculating. Hopefully this is because they're tracking and allowing for many moving parts, and not because they are Excel sheets using vBasic macros written by over-worked, under-paid, self-taught office clerks.

This hits very close to home. I was that over-worked, under-paid, self-taught vBasic programmer. Well, retired now. But we got a lot of work out of those macros, processing the output of an automated test setup and crunching data from multiple sources into a human-readable chart that saved so much time from them doing it manually that I got a promotion (I saved them a lot of time - like a day and a half's work got done in ten minutes). But I never programmed a system to jump.

 

[Condottiere]

Artificial intelligence programmed programmes don't tend to be elegant.

But inexpensive.

 

[Grav_Moped]

The computers have to mine enough Bitcoin to pay for Chat-GPT to plot the Jump course.

The unusuall size is because they include a power source for that (and it's never enough anyhow).

 

[Condottiere]

That would make that a jumpway toll.

 

[Rupert]

This hits very close to home. I was that over-worked, under-paid, self-taught vBasic programmer. Well, retired now. But we got a lot of work out of those macros, processing the output of an automated test setup and crunching data from multiple sources into a human-readable chart that saved so much time from them doing it manually that I got a promotion (I saved them a lot of time - like a day and a half's work got done in ten minutes). But I never programmed a system to jump.

The comment was meant as humour, and I apologise if it caused any offence. I did a little myself, in the day. The code I produced probably varied from acceptable (that which was simple hacks on someone else's original work) to 'well it works - mostly'. Mainly I was referring to those insane excel sheets that seem to accrete 'stuff' as they get passed back and forth in some workplaces, and which choke up quite powerful PCs any time you try and do anything with them.

 

[Badenov]

The comment was meant as humour, and I apologise if it caused any offence. I did a little myself, in the day. The code I produced probably varied from acceptable (that which was simple hacks on someone else's original work) to 'well it works - mostly'. Mainly I was referring to those insane excel sheets that seem to accrete 'stuff' as they get passed back and forth in some workplaces, and which choke up quite powerful PCs any time you try and do anything with them.

Oh, no offense taken. I was amused, as much as anything, to imagine I had a place in the Traveller universe, if not doing my primary job. But clearly it's an important job, if it can save so much work for those not quite up to speed. I've seen the types of accumulated data bases you're speaking of, and they're invariably handled by people who got their spreadsheet skills from a cereal box. If you could write macros for that and they work at all, you are pretty talented, yourself. I wrote absolute spaghetti, myself, but confident that there was no one anywhere close in my organization who could so much as write a comment in it, and as long as it worked, I was golden.

But to tie this back to the thread, I will pose the question: how complex is a TL5 computer that is far less hefty than ENIAC and other computers of that generation. Is it like one of the mechanical computers the Navy used for fire control through the 1960's? Those were amazingly complex for what they did, but not much compared to their electronic successors at later TLs. At 1 dT, it's far closer to the size of one of those mechanical computers than to ENIAC.

 

[kilemall]

Oh, no offense taken. I was amused, as much as anything, to imagine I had a place in the Traveller universe, if not doing my primary job. But clearly it's an important job, if it can save so much work for those not quite up to speed. I've seen the types of accumulated data bases you're speaking of, and they're invariably handled by people who got their spreadsheet skills from a cereal box. If you could write macros for that and they work at all, you are pretty talented, yourself. I wrote absolute spaghetti, myself, but confident that there was no one anywhere close in my organization who could so much as write a comment in it, and as long as it worked, I was golden.

But to tie this back to the thread, I will pose the question: how complex is a TL5 computer that is far less hefty than ENIAC and other computers of that generation. Is it like one of the mechanical computers the Navy used for fire control through the 1960's? Those were amazingly complex for what they did, but not much compared to their electronic successors at later TLs. At 1 dT, it's far closer to the size of one of those mechanical computers than to ENIAC.

I’m pretty convinced Miller was modeling the USN computers, initially the mechanical ones on the battleships and later the digital ones.

 

[Rupert]

At TL5 those old 'range tables' and 'Mk1 fire control computers' are probably it, though apparently with some for of programmability, which suggests some kind of re-wire-able relay or valve array or the like - somehow I doubt 'programs' are large modules of complicated clockwork (though that'd be cool in it's own way that's not supported by the rules or fluff).

TL6 is valve-technology and TL7 is transistor/early IC tech, so we've got clear examples of how computers at those TLs were expected by the authors to be built.

 

[Badenov]

I’m pretty convinced Miller was modeling the USN computers, initially the mechanical ones on the battleships and later the digital ones.

This is another bump for computer rating being part of your to-hit roll in LBB5. It seems like the shipboard computer's primary job is fire control, and jump control is a secondary feature.

No, but you have to solve hyperdimensional general relativistic n-body problems to plot the jump. The system you are in is an n-body problem, the system you are jumping to is an n-body problem (based on a model of that system since the data for the position and velocity of every object in that system is several years out of date and has to be modeled - ask your AI of choice how far into the future our best models of our solar system are good for ).

So maybe this is why there's the 100-diameter limit, outside which these n-bodies reduce to negligibility and allow much simpler calculation. It's jumping inside the limit where you definitely need to juggle n-body math, which apparently they still can't do at TL15.

It might be interesting if very high TL computers could jump closer than 100 diameters.