Revamp of ships' computers
- Thread starter JAFARR
- Start date
TheEngineer
SOC-14 1K
Hi folks !
Just to give some inspiration about stats of massively parallel working supercomputers...
Real World Supercomputers
There is a bunch of sub-links to actual supercomputing sites.
Another good idea is to do visit to the next nuclear powerplant, and fetch some information about the stats of the computer systems used there to control the business.
Perhaps this might give another impression, that there is another dimension of computing (which in fact is unknown to the most "IT-professionals", becauses the sites are staffed with engineers or scientists), turning even Traveller computers into somehow cute installations.
So, if you combine the knowledge about process control computing for a complex technical installation (power plant, refinary, automotive) with the dimensions of real world supercomputers, there shouldnt be a problem left with the Traveller computer stuff.
And obviously the jump capability is the thing, which really needs the massive parallel realtime supercomputing capabilities, resulting in those vast volumes even at TL15.
As a result I wouldnt see any reason to tell, that there is something wrong with the rules
Regards,
Mert
Just to give some inspiration about stats of massively parallel working supercomputers...
Real World Supercomputers
There is a bunch of sub-links to actual supercomputing sites.
Another good idea is to do visit to the next nuclear powerplant, and fetch some information about the stats of the computer systems used there to control the business.
Perhaps this might give another impression, that there is another dimension of computing (which in fact is unknown to the most "IT-professionals", becauses the sites are staffed with engineers or scientists), turning even Traveller computers into somehow cute installations.
So, if you combine the knowledge about process control computing for a complex technical installation (power plant, refinary, automotive) with the dimensions of real world supercomputers, there shouldnt be a problem left with the Traveller computer stuff.
And obviously the jump capability is the thing, which really needs the massive parallel realtime supercomputing capabilities, resulting in those vast volumes even at TL15.
As a result I wouldnt see any reason to tell, that there is something wrong with the rules
Regards,
Mert
I'll delurk briefly to offer support for Uncle Bob's take on this topic (and yes, IT pros still play Traveller 
).
I've struggled with how I wanted to change the computer technology IMTU, particularly with regard to starships, and got caught up in the whole idea of volume/power/capacity curves. Reading Uncle Bob's proposal in this thread, the weight has been lifted from my shoulders and has lit the bulb over my head -- now I finally see the light. Praise Bob!
I think the original Book 2 comps are loading the "jump tapes" off 8-track cartridges (I'm old enough to have been using 8-track when I first played Traveller).
Which is kind of a hard-sell to my current players, much less my own taste. I toyed with idea of using big analog vacuum computers based on ceramic plates (an idea pitched years ago on the TML) as a way to avoid EMP-like complications of jump space transfer, but optical systems would ultimately solve that problem long before TL15. I was never really happy with any of the solutions I've seen in later editions, either.
The answer is, its all about the software. Above a certain tech level (TL9 at most, I'd say), the computer hardware becomes completely commoditized, distributed, and transparent. Our materials technology today is able to imprint microscopic circuitry in clothing, by TL13 (Imperial average tech) embedded nanoscopic optical systems should have vast computational abilities and virtually infinite storage. The data systems on a TL13 ship's boat could store the entire sum knowledge of the Imperium (not that it would, of course
).
I realize the implications of this depart from canon in the direction of something akin to Dan Simmon's Hyperion. On the otherhand, I'm a flaming heretic when it comes to some issues, and I'm looking for alternatives to a canonized hairshirt when it comes to computers.
There was an network engineer on a messageboard I read years ago whose signature tag read something like, "Unless its on fire, its a software problem." I think this will be even more true in the future, and also points the area of real concern when it comes to starship data systems. While computational hardware may reach a functional horizon relatively early in the tech tree, this does not mean that software technology follows the same path. While storage size issues may become moot, fictional contructs such as AI and jumpspace mathematics can still be constrained to the contantly emerging technology curve of my choice -- just because hardware limitations cease does not make such advanced software technology inevitable, much less free.
To sum: IMTU starship computer hardware (as such) is subsumed into the cost, tonnage, and energy requirement of all data controlled components (bridge, comms, senors, drives, etc.), and there are no practical limits on data storage capacity. All variables of cost, capability, and performance are software dependent. If the Tech Level of the central data processing hardware ever comes into question, the TL of the ship's bridge construction shall be used.
Heresy, I know. I do enjoy the notion of Hobart Floyt transferring his favorite AI into his pocketwatch, or Cortana (the ship's AI from my namesake video game) downloading herself onto a business card sized chip. Even if I do limit that sort of thing to TL15+.
).I've struggled with how I wanted to change the computer technology IMTU, particularly with regard to starships, and got caught up in the whole idea of volume/power/capacity curves. Reading Uncle Bob's proposal in this thread, the weight has been lifted from my shoulders and has lit the bulb over my head -- now I finally see the light. Praise Bob!
I think the original Book 2 comps are loading the "jump tapes" off 8-track cartridges (I'm old enough to have been using 8-track when I first played Traveller).
The answer is, its all about the software. Above a certain tech level (TL9 at most, I'd say), the computer hardware becomes completely commoditized, distributed, and transparent. Our materials technology today is able to imprint microscopic circuitry in clothing, by TL13 (Imperial average tech) embedded nanoscopic optical systems should have vast computational abilities and virtually infinite storage. The data systems on a TL13 ship's boat could store the entire sum knowledge of the Imperium (not that it would, of course
I realize the implications of this depart from canon in the direction of something akin to Dan Simmon's Hyperion. On the otherhand, I'm a flaming heretic when it comes to some issues, and I'm looking for alternatives to a canonized hairshirt when it comes to computers.
There was an network engineer on a messageboard I read years ago whose signature tag read something like, "Unless its on fire, its a software problem." I think this will be even more true in the future, and also points the area of real concern when it comes to starship data systems. While computational hardware may reach a functional horizon relatively early in the tech tree, this does not mean that software technology follows the same path. While storage size issues may become moot, fictional contructs such as AI and jumpspace mathematics can still be constrained to the contantly emerging technology curve of my choice -- just because hardware limitations cease does not make such advanced software technology inevitable, much less free.
To sum: IMTU starship computer hardware (as such) is subsumed into the cost, tonnage, and energy requirement of all data controlled components (bridge, comms, senors, drives, etc.), and there are no practical limits on data storage capacity. All variables of cost, capability, and performance are software dependent. If the Tech Level of the central data processing hardware ever comes into question, the TL of the ship's bridge construction shall be used.
Heresy, I know. I do enjoy the notion of Hobart Floyt transferring his favorite AI into his pocketwatch, or Cortana (the ship's AI from my namesake video game) downloading herself onto a business card sized chip. Even if I do limit that sort of thing to TL15+.
Except, commercial nuclear power plants are all at least thirty years old, and the designs are 40. Military power plants are too classified for me to comment on, but many designs are almost as old. And most of a state-of-the art process control center is volume we would call "bridge". At any rate, the ship's computer has no function whatsoever with ships systems. Proved by the fact that bigger and more sophisticated systems do not demand bigger computers. Also, damage to the computer does not disable any other system.Originally posted by TheEngineer:
So, if you combine the knowledge about process control computing for a complex technical installation (power plant, refinary, automotive) with the dimensions of real world supercomputers, there shouldnt be a problem left with the Traveller computer stuff.
And obviously the jump capability is the thing, which really needs the massive parallel realtime supercomputing capabilities, resulting in those vast volumes even at TL15.
As a result I wouldnt see any reason to tell, that there is something wrong with the rules
Regards,
Mert
Book 2 clearly shows you don't need a supercomputer to jump. It takes roughly the same "space" as normal navigation, which can be done at TL6 on a PDP-11, or at TL-7 by my old Commaore-64.
Your arguments did not stand up when I made them in 1980, and they still don't. They have just been getting stale for a quarter century.
I am sure I have been through this before. Ten or twelve times.
The computer rules are broken.
Except, commercial nuclear power plants are all at least thirty years old, and the designs are 40. Military power plants are too classified for me to comment on, but many designs are almost as old. And most of a state-of-the art process control center is volume we would call "bridge". At any rate, the ship's computer has no function whatsoever with ships systems. Proved by the fact that bigger and more sophisticated systems do not demand bigger computers. Also, damage to the computer does not disable any other system.Originally posted by TheEngineer:
So, if you combine the knowledge about process control computing for a complex technical installation (power plant, refinary, automotive) with the dimensions of real world supercomputers, there shouldnt be a problem left with the Traveller computer stuff.
And obviously the jump capability is the thing, which really needs the massive parallel realtime supercomputing capabilities, resulting in those vast volumes even at TL15.
As a result I wouldnt see any reason to tell, that there is something wrong with the rules
Regards,
Mert
Book 2 clearly shows you don't need a supercomputer to jump. It takes roughly the same "space" as normal navigation, which can be done at TL6 on a PDP-11, or at TL-7 by my old Commaore-64.
Your arguments did not stand up when I made them in 1980, and they still don't. They have just been getting stale for a quarter century.
I am sure I have been through this before. Ten or twelve times.
The computer rules are broken.
Except, commercial nuclear power plants are all at least thirty years old, and the designs are 40. Military power plants are too classified for me to comment on, but many designs are almost as old. And most of a state-of-the art process control center is volume we would call "bridge". At any rate, the ship's computer has no function whatsoever with ships systems. Proved by the fact that bigger and more sophisticated systems do not demand bigger computers. Also, damage to the computer does not disable any other system.Originally posted by TheEngineer:
So, if you combine the knowledge about process control computing for a complex technical installation (power plant, refinary, automotive) with the dimensions of real world supercomputers, there shouldnt be a problem left with the Traveller computer stuff.
And obviously the jump capability is the thing, which really needs the massive parallel realtime supercomputing capabilities, resulting in those vast volumes even at TL15.
As a result I wouldnt see any reason to tell, that there is something wrong with the rules
Regards,
Mert
Book 2 clearly shows you don't need a supercomputer to jump. It takes roughly the same "space" as normal navigation, which can be done at TL6 on a PDP-11, or at TL-7 by my old Commaore-64. We don't know how to do it yet, but when we figure it out it won't be hard to do
Your arguments did not stand up when I made them in 1980, and they still don't. They have just been getting stale for a quarter century.
I am sure I have been through this before. Ten or twelve times.
The computer rules are broken.
Except, commercial nuclear power plants are all at least thirty years old, and the designs are 40. Military power plants are too classified for me to comment on, but many designs are almost as old. And most of a state-of-the art process control center is volume we would call "bridge". At any rate, the ship's computer has no function whatsoever with ships systems. Proved by the fact that bigger and more sophisticated systems do not demand bigger computers. Also, damage to the computer does not disable any other system.Originally posted by TheEngineer:
So, if you combine the knowledge about process control computing for a complex technical installation (power plant, refinary, automotive) with the dimensions of real world supercomputers, there shouldnt be a problem left with the Traveller computer stuff.
And obviously the jump capability is the thing, which really needs the massive parallel realtime supercomputing capabilities, resulting in those vast volumes even at TL15.
As a result I wouldnt see any reason to tell, that there is something wrong with the rules
Regards,
Mert
Book 2 clearly shows you don't need a supercomputer to jump. It takes roughly the same "space" as normal navigation, which can be done at TL6 on a PDP-11, or at TL-7 by my old Commaore-64.
Your arguments did not stand up when I made them in 1980, and they still don't. They have just been getting stale for a quarter century.
I am sure I have been through this before. Ten or twelve times.
The computer rules are broken.
Fritz_Brown
Super Moderator
Dude, so is your computer! Or your connection. 4 simultaneous posts - that is a new record, I think!
Hardly. http://www.johnstonsarchive.net/environment/nuclearplants.html
Errr... yes they do. We are talking about CT/HG, right?
Yes, it does. Apart from the fact that in HG, damage to the computer severely affects offensive and defensive power, you need a computer to jump.
Regards,
Tobias
TheEngineer
SOC-14 1K
Ok. Peace
Just to clarify: if I talk about Traveller computers I talk about the way they where presented in MT. Computer rules are more abstract here, and except the statement of the enzyclopedia, that a TL 11 hand computer is analogue to a supercomputer or a model/1, there is no actual performance comparision to "real life". (as there is a fictional performance needed to run fictional devices, it might be hard to say there is something wrong here).
So, CT or book 2 stuff is perhaps something different and I would consider those rules as surely "outdated" and partly "broken", too (e.g. the program size and load capacities)...
Regarding nuclear plants I would like to ask, if somebody knows if US sites are forced to follow technical advancement as close as this is done here in Germany (they always have to use "state of advanced techniques" here, if security aspects are concerned) ?
Best regards,
Mert
Just to clarify: if I talk about Traveller computers I talk about the way they where presented in MT. Computer rules are more abstract here, and except the statement of the enzyclopedia, that a TL 11 hand computer is analogue to a supercomputer or a model/1, there is no actual performance comparision to "real life". (as there is a fictional performance needed to run fictional devices, it might be hard to say there is something wrong here).
So, CT or book 2 stuff is perhaps something different and I would consider those rules as surely "outdated" and partly "broken", too (e.g. the program size and load capacities)...
Regarding nuclear plants I would like to ask, if somebody knows if US sites are forced to follow technical advancement as close as this is done here in Germany (they always have to use "state of advanced techniques" here, if security aspects are concerned) ?
Best regards,
Mert
Okay, if we are talking about book 2, then I will agree that computers are broken badly. I mean, the "tape" thing alone. But HG abstracts computers enough that they can be interpreted as "fitting", by including sensors, comm, etc.
An SF game should never make the mistake of rating its computers in real-life stats, and generally should strife to leave this particular technology somewhat fuzzy, because it happens to change so quickly in our times. Similar if someone in 1910 had published an SF RPG with airplanes.
Regards,
Tobias
An SF game should never make the mistake of rating its computers in real-life stats, and generally should strife to leave this particular technology somewhat fuzzy, because it happens to change so quickly in our times. Similar if someone in 1910 had published an SF RPG with airplanes.
Regards,
Tobias
Andrew Boulton
The Adminator
There are some interesting aircraft predictions from 1916 here: http://ansernet.rcls.org/lawson/prim3b.htm
Sorry about the multiple posts
Originally posted by Tobias:
I stand corrected.
There is a relationship to overall ship size, not to the complexity of the systems. More like the way Lloyds and the USCG require larger ships to have radars.
As Ihave always said, it affects navigation and targetting, only. The airconditioning keeps humming and the power planrt keeps glowing. And the weapons hit just as hard, so their function is not "ship's computer" dependent, just their aim.
Originally posted by Tobias:
Hardly. http://www.johnstonsarchive.net/environment/nuclearplants.html
I stand corrected.
Errr... yes they do. We are talking about CT/HG, right0?
There is a relationship to overall ship size, not to the complexity of the systems. More like the way Lloyds and the USCG require larger ships to have radars.
Yes, it does. Apart from the fact that in HG, damage to the computer severely affects offensive and defensive power, you need a computer to jump.
As Ihave always said, it affects navigation and targetting, only. The airconditioning keeps humming and the power planrt keeps glowing. And the weapons hit just as hard, so their function is not "ship's computer" dependent, just their aim.
I don't want to jump into an argument either, but I disagree that Book 2 computers have serious problems.
I have book2 and the Traveller book and the only reference to computer media is the self-erasing cassette for navigation flight plans. I don't see where it says "magnetic tape" and even if it is, why is that the only media available?
I like the programming rules. I may be a minority but I think it makes combat more interesting because you have more choices to make.
The small craft rules sort of imply that there are other computers on a ship. A small craft doesn't need a "computer" all it needs is a bridge. It seems that there are a number of computers included in the bridge volume that control power plant, housekeeping, drives and so on. The big box is used to tie programs together (maneuver/evade) and to run high speed simulations (predict, anti-hijack).
Just my two cents.
I have book2 and the Traveller book and the only reference to computer media is the self-erasing cassette for navigation flight plans. I don't see where it says "magnetic tape" and even if it is, why is that the only media available?
I like the programming rules. I may be a minority but I think it makes combat more interesting because you have more choices to make.
The small craft rules sort of imply that there are other computers on a ship. A small craft doesn't need a "computer" all it needs is a bridge. It seems that there are a number of computers included in the bridge volume that control power plant, housekeeping, drives and so on. The big box is used to tie programs together (maneuver/evade) and to run high speed simulations (predict, anti-hijack).
Just my two cents.
Well someone who reads Niven can't be all wrong!
No where does it say magnetic tape - correct as far as it goes. However, there has to be a way to transport software from computer to computer. To continue with the line of reasoning and emply that the rules mean you un-install one program, install another and use it to perform an operation of some sort ALL in one segment of a 15 minute turn... Maybe they intended something like the old ROM pack games for the early computers that you connected to your TV.
Sorry, but something is broken and has been the whole time, but nobody seems to be able to agree on a workable solution.
