Computer usage
- Thread starter Willie Dewitt
- Start date
That's a pretty fair point. If you want to ruggardise anything to make it properly milspec, then it's going to be bigger or at least bulkier. Then there's cost. That doesn't mean we have to imagine them as 1970's machines in space, unless that's the trope you're running with in your game. But the point that it's running a fusion power plant, grav and inertial compensation system, life control, the myriad of ship systems, all while having redundancy built in so there's not catastrophic failures (there's a use for the T5 QREBS system with your group's 200 year old TL10 Free Trader which they've been skimping on service costs for) well points to them being at least moderately sophisticated machines.
Let my team have an expert system with a personality help run their ship. Not quite Holly, but fun to play nonetheless, and it allowed me to have a bit more inject into their problem solving in the most sarcastic manner possible.
kilemall
SOC-14 5K
Oh ya, on board with what works. If you look at my links, I have the cheaper machines, but you lose passenger certification and increasing unreliability and maintenance issues the cheaper you go.
That's directly from my experience as the computer world went from first class mint on your pillow service with IBM to 'here's your cracker and soda shuddup and stay in your seat' crowded coach class of today.
I also had a separate proposed thread on putting the INT in your INTEL machine, basically INT level per model, entirely deterministic machine as you don't need HAL/sulking ship computer issues in the middle of trouble and the INT is just translating commands into actions the computer can take with equipment it's connected to.
But yes for the price one should get more value for all those credits, seems a reasonable way to do that and make one feel the difference between TL9 tramp freighters and a sleek TL14 exploration vessel.
Actually there could be another "viable" reason for the size / weight of the CT shipboard computers, particularly in older and exploratory classes of vessels. They could be being built using "obsolete / obsolescent" technology like discrete components and printed circuit boards that can be readily repaired by anyone with some basic electronics knowledge, a steady hand and a soldering iron (or a red hot nail in a pinch 
) After all, if you're stranded on an uncharted planet in the middle of nowhere after a misjump with a computer failure, would you want to hear the ship's engineer say....
"I'm sorry folks, I don't have a spare XYZ Industries computer power supply module to swap for the faulty one, guess we're here for the duration...."
or
"No worries, I've traced the fault to a blown voltage regulator in the power supply. Give me 20 minutes to cannibalize one from the coffee maker and Nav can start plotting our way home.....what? Why 20 minutes when it's a 5 minute job? Well, if I'm going to cannibalize the coffee maker you can be sure I'm going to get one last cup brewed first!" :rofl:
) After all, if you're stranded on an uncharted planet in the middle of nowhere after a misjump with a computer failure, would you want to hear the ship's engineer say...."I'm sorry folks, I don't have a spare XYZ Industries computer power supply module to swap for the faulty one, guess we're here for the duration...."
or
"No worries, I've traced the fault to a blown voltage regulator in the power supply. Give me 20 minutes to cannibalize one from the coffee maker and Nav can start plotting our way home.....what? Why 20 minutes when it's a 5 minute job? Well, if I'm going to cannibalize the coffee maker you can be sure I'm going to get one last cup brewed first!" :rofl:
Timerover51
SOC-14 5K
A Model 7 Computer is listed at 9 Traveller Displacement Tons, or just under 4500 cubic feet. That is a room 10 feet high by 15 feet wide by 30 feet long.
A ship bridge is defined at 2 per cent of a ship's tonnage, with a minimum of 20 Traveller Displacement tons of space required. On a ship of less than or equal to 1000 tons, the Model 7 computer is close to one-half the size of the ship's bridge.
As for power plant control, on 17 January 1955, the USS Nautilus signaled that she was "Underway on nuclear power." At best, her computer would be rated a Model 1 or maybe Model 1/bis. Power plant control is going to be in the Engineering area, and included in the cost of the engineering plant. If you want redundancy, you have three computers operating, one main and two back-up. The same would be true of your bridge computer.
I fail to see a need for allotting 500 cubic feet of space for a Tech Level 8 Model 1 computer, or even three of them, given the current Real World state of the art. I most definitely fail to see the need to allot 4500 cubic feet of volume for a Model 7.
Ship's computer volume should be included in the 20 Traveller Displacement tons of Bridge space, power plant computer operation should be included in the engineering spaces, and if you wish to add the additional cost to the engineering department, feel free to do so.
A ship bridge is defined at 2 per cent of a ship's tonnage, with a minimum of 20 Traveller Displacement tons of space required. On a ship of less than or equal to 1000 tons, the Model 7 computer is close to one-half the size of the ship's bridge.
As for power plant control, on 17 January 1955, the USS Nautilus signaled that she was "Underway on nuclear power." At best, her computer would be rated a Model 1 or maybe Model 1/bis. Power plant control is going to be in the Engineering area, and included in the cost of the engineering plant. If you want redundancy, you have three computers operating, one main and two back-up. The same would be true of your bridge computer.
I fail to see a need for allotting 500 cubic feet of space for a Tech Level 8 Model 1 computer, or even three of them, given the current Real World state of the art. I most definitely fail to see the need to allot 4500 cubic feet of volume for a Model 7.
Ship's computer volume should be included in the 20 Traveller Displacement tons of Bridge space, power plant computer operation should be included in the engineering spaces, and if you wish to add the additional cost to the engineering department, feel free to do so.
mike wightman
SOC-14 10K
A model 7 computer is TL13 - probably a lot more likely to be some sort of quantum computer way beyond our current understanding.
Think about what a ship computer has to be.
A computer that can run a hyperdimensional n-body problem in real time would be at minimum the size of a supercomputer today - and that is a model 1.
Add to that it is controlling an artificial gravity field, acceleration compensation, full environmental control, a fusion reactor, avionics, sensors and possible tetris all at the same time.
You are talking supercomputer, not desktop or laptop, and supercomputers take up hundreds of square metres of floor space and require megawatts of energy.
Think about what a ship computer has to be.
A computer that can run a hyperdimensional n-body problem in real time would be at minimum the size of a supercomputer today - and that is a model 1.
Add to that it is controlling an artificial gravity field, acceleration compensation, full environmental control, a fusion reactor, avionics, sensors and possible tetris all at the same time.
You are talking supercomputer, not desktop or laptop, and supercomputers take up hundreds of square metres of floor space and require megawatts of energy.
Last edited:
kilemall
SOC-14 5K
I wasn't suggesting that the ship's computer actually ran the power plant, more like coordinated it with all the other involved major subsystems, most notably the Maneuver or Jump programs and associated drives.
Presumably when one executes a violent Evade manuever, the ship's computer will need to know that in order to adjust Target and Gunnery/Predict tracking to maintain an optimal shot.
As for size of a Model/7 on a smallish craft, I absolutely DO like that size, because to me unless it is running some of those HG-sized ships, that computer is there to run the TL13 equivalent of Aegis Fire Control/AWACS systems along with an EW capability (given the HG degrade to lower models against it), and that level of performance does NOT fit in a server rack.
Want to splurge on unlimited computing Predict, multi-track, gunnery, maneuver/evade power, etc., or get the HG +7 DM, got to pay the piper. Just on gaming balance principles alone it's a near must, as computers either CT or HG are ship power multipliers.
As for floor space, I'm good with a distributed computer all over the ship that degrades gracefully and 1-2 master consoles rather then all in one room- fits the CT 'damn hard to kill' computer damage model.
Presumably when one executes a violent Evade manuever, the ship's computer will need to know that in order to adjust Target and Gunnery/Predict tracking to maintain an optimal shot.
As for size of a Model/7 on a smallish craft, I absolutely DO like that size, because to me unless it is running some of those HG-sized ships, that computer is there to run the TL13 equivalent of Aegis Fire Control/AWACS systems along with an EW capability (given the HG degrade to lower models against it), and that level of performance does NOT fit in a server rack.
Want to splurge on unlimited computing Predict, multi-track, gunnery, maneuver/evade power, etc., or get the HG +7 DM, got to pay the piper. Just on gaming balance principles alone it's a near must, as computers either CT or HG are ship power multipliers.
As for floor space, I'm good with a distributed computer all over the ship that degrades gracefully and 1-2 master consoles rather then all in one room- fits the CT 'damn hard to kill' computer damage model.
kilemall
SOC-14 5K
Mike, he's got a point about the control-electronics built into the different ship systems, each probably has it's own computer specialized for realtime engineering control. They just aren't multi-purpose or linked systems designed to 'put it all together' for the ship's pilot, navigator/jump engineers and/or gunners.
I don't know in that case that you NEED a supercomputer per se, a lot of the controls and handling of the various calcs needed to operate the equipment would be built in, the ship's computer would be more the central point at which all these systems come together and feed each other the data they need to adjust to the current crew's actions.
I don't know in that case that you NEED a supercomputer per se, a lot of the controls and handling of the various calcs needed to operate the equipment would be built in, the ship's computer would be more the central point at which all these systems come together and feed each other the data they need to adjust to the current crew's actions.
Timerover51
SOC-14 5K
kilemall
SOC-14 5K
No, plenty of room for those subsector space control type smallish warships that are nth better then the corsairs or larger brothers but not mainline units.
Also spreads the logistical load of shipyard support to sustainable levels, spreads the wealth around to worlds less then A-TL F and stretches the taxpayer's credit to get the local jobs done and have plenty for the Big Guns.
However, I think there is room for high speed light recon cruisers that have top fitted sensor suites for fleet scouting- big enough to sweep away enemy light scouting units and fast enough to run from mainline units. If computer models gives me game shorthand for expending extra cost for that extra capability and I don't have to melt my TI calculator in the process, well good enough.
NO thanks, I'm good with the big expensive computer models.
As a matter of fact, in the Sensors and Engagement Ranges thread and earlier ones, I've stated that yes the entire hull is a sensor grid, for the same reasons as the distributed computer, damage does not equal capability loss. Power and control cabling, life support, etc. get the same treatment.
I'm grappling with the sensor ranges as part of the desired game effect, wanting a simple system that doesn't take Striker levels of design work to deploy but 'feels' complex and 'right' and differentiated between TLs and invested computing and fits into the 'right' interaction between maneuver and fire and ship design. Posted it there if you want to see it.
I'm not out to put real world numbers on it, but if you take the 4 MW of an AN/SPY-1 and convert that to HG power, that's 8 EP. A model/7 is 7 EP, so eh, in the ballpark, if it makes you feel better.
Sorry OP, we ARE getting into some esoterica, but that's what happens in gearheadland. Most of this is not applicable to pure CT computer wrangling.
Did you find my example post useful, any clarifications I should make?
kilemall
SOC-14 5K
And just for giggles, what IF we actually built that TL5 analog Model/1?
Well, it would probably be THIS-
https://www.youtube.com/watch?v=s1i-dnAH9Y4
Hehehe, one of these days I intend to strand players out in the Oort cloud or on a low tech planet, and FORCE an analog gear computer on them, because that's all they can make Out There, and it keeps the ice ships moving.
Hey. Invulnerable to rad hits. That counts.
Well, it would probably be THIS-
https://www.youtube.com/watch?v=s1i-dnAH9Y4
Hehehe, one of these days I intend to strand players out in the Oort cloud or on a low tech planet, and FORCE an analog gear computer on them, because that's all they can make Out There, and it keeps the ice ships moving.
Hey. Invulnerable to rad hits. That counts.
mike wightman
SOC-14 10K
The target program alone allows for tracking and shooting stuff at ranges measured in light seconds, that's a lot better than anything we have today.
It has to be in order to explain the sensor ranges and capabilities of CT ships.
mike wightman
SOC-14 10K
Oh I agree that there will be control electronics built into most stuff on board the ship, but the main computer gets the job of making sure everything works together.
The reason I have always thought super computer is the real time calculation of jump solutions - the generate program. I can't imagine anything but a super computer being able to cope with those calculation.
We hit a 1m square weekly at 1.5LS at present. At better than 99.99% accuracy. True, it's movement isn't that big, and it's at a known point on the moon... but we still have to lead the target. Which said target does 3,683 kilometers per hour, or better than 1 kps... and we have to aim based upon 1.25 sec old data for a future position 1.25 sec away, so, just shy of 3km off target's present known position (which is 1.25 sec out of date) and the beam is hitting with a beam that, with the telescope position off by 1700m, missed the retroreflectors... which means the beam width at target (1.25 ls away) is probably less than 1700m.
We had that accuracy in August 1969. It's gotten better since then.
CIWS has accuracy better than 1cm at 10km with projectiles in air and a 99% hit rate...
Accuracy isn't the issue the computer has to solve. (The needed calcs can now be done on high end digital watches!)
http://coolcosmos.ipac.caltech.edu/ask/176--How-fast-does-the-Moon-travel-around-Earth-
https://www.ucolick.org/news/2009/apollo11.pdf
kilemall
SOC-14 5K
This goes in hand with my assertion that virtually every Traveller beam shot is a 'hit', the question is whether it's a solid enough hit on the side of the craft with enough energy to penetrate and not an edge hit or a shot that had to spread out it's fire due to too large a potential endpoint set.
mike wightman
SOC-14 10K
I don't think hitting a target on the moon from a fixed emplacement on the Earth is remotely similar to ship to ship combat. :devil:
Hitting an evading target from a platform that is itself trying to avoid being hit is a bit different. If such weapon systems were possible then ballistic missile defence would be trivial. As would active defence vs tank rounds, rpgs etc.
Hitting an evading target from a platform that is itself trying to avoid being hit is a bit different. If such weapon systems were possible then ballistic missile defence would be trivial. As would active defence vs tank rounds, rpgs etc.
Your conflating two very different issues...
Accuracy of beam
collimation of beam
We can put any beam we can generate on that kind of accuracy.
What we cannot do at present is get a collimation down enough that energy density remains high enough to melt steel past about 1.5 km. Partly, that's due to atmosphere effects, partly due to the lasing apparatus.
We routinely point telescopes to ≤0°0'4" from a dirtside scope of some age (keck).
We routinely get sensor accuracy of under an arcsecond with CIWS, and pretty close to same in putting bullets on target. 2 cm at 1.5km published accuracy; evidentiary of 10cm at 10km. 0.01:10000, or 1:1,000,000 ... which is, at present, pretty much going to be more accurate than we can focus.
Meh, the math is the same -- it's the hardware that makes the difference. It's really just simple ballistics with the domain of being in space being considered rather than tracking wind and shell drift like we have here on earth.
kilemall
SOC-14 5K
Well it's always able to have been done- from the 70s on- just a matter if you have the money and will to fund the genius and engineering.
Read this one where they were tearing out cards and turning the power off for the Safeguard ABM computers-
http://www.nuclearabms.info/Computers.html
