• Welcome! The TrekBBS is the number one place to chat about Star Trek with like-minded fans.
    If you are not already a member then please register an account and join in the discussion!

The computers of Star Trek

knightgrace

Captain
Captain
I was recently rereading the first chapters in this book.

ENIAC had 18,000 vacuum tubes. Eighteen thousand! But contained in this is a different way of looking at ENIAC.

AS 18,000 switches, or 18,000 bits of data.

Using Steve Long's Space Dock Original Series Sourcebook game supplement a Constitution class starship main computer cores, each contains 8,500 kiloquads of data. Or to put it another way, 900,000+ Duotronic relays. Actually 901,000, to divide evenly - 106 switches per quad.

A relay, however doesn't need to be a simple on/off...

It can be multifunction.

But! Converting back, this means that the Enterprise computer core would be 106 times as powerful as a an IBM 360.
So, there you have it.

Exactly how much computer power was figured for Star Trek.

Keeping in mind that a relay isn't a simple switch. This non simple device explains a great deal.

By the way, relay logic is usually implemented as ladder logic.
 
And Data is as smart as 2 top of the line Nvidia GPU's, or something. Back when we started watching Trek as kids, touchscreens were the future. But they were 30 years in the future not 300. I suspect the rig running the on-set graphics in modern Trek is more powerful than how they imagined future computers in TOS.
 
In Omni magazine, back in the the early 1990s, ran an article on the upcoming Star Trek: Voyager series, and the article said that one of the creators of Voyager knew that the computer concepts being designed was only fifteen years at most...

In other words, Voyager was under powered, computer wise.

Which brings up the question of just how A. I. Voyager actually would be.

Hint: Ship in the Bottle, but with far more EHs.
 
And Data is as smart as 2 top of the line Nvidia GPU's, or something. Back when we started watching Trek as kids, touchscreens were the future. But they were 30 years in the future not 300. I suspect the rig running the on-set graphics in modern Trek is more powerful than how they imagined future computers in TOS.
One thing I like to point out is that all the TOS behind-the-scenes material referred to Spock's station not as "science" by as the Library Computer station. It was implicitly the only place on the bridge that could actually pull information out of the ship's mainframe.

It's a small wonder they started referring to all their computer speeds and feeds as being measured in "quads." Using a real-world figure like operations-per-second for Data was a surprisingly late flub (though, if you wanted to fanwank it, you could say an "operation" in 24th century computer science isn't the same as what it is today). I'm also reminded of the first episode of the BSG prequel "Caprica," which initially referenced the amount of data needed to make a convincing digital avatar of a person being in the hundreds of megabytes, which was changed to be on the order of terabytes for the final aired version, based on a contemporary estimate of the size of the human brain translated into bits and bytes. I'm not sure that was actually necessary, IIRC the conversation was talking about how much personal data needed to be harvested from the internet to acceptably (by some standard) emulate a person's personality, not the actual personality itself, and a few hundred megabytes of text might not be off-base to build a personality profile.
 
Three hundred megabytes. Yes, I was initially startled by this figure.


Consider for production system, original called an expert system, needs eight hundred if then else rules to be moderately competent...

So, three hundred megabytes isn't too far fetched, for completeness. Not including image data.

What Zoe Graystone did, is vastly out perform her father, which was the point. He couldn't figure it out, she did.

And I haven't even gotten into fuzzy logic yet, which basically multiplies each if then else statement by a factor of ten. (This changes over time if the fuzzy logic weights can change).

What is implied by the quad, which is beyond binary, is that Metadata is, while connected to the data, is processed separately from the straight line data.

In Caprica, the competition against Grayson Industries developed the metacognitive processor. This is obviously a place holder term, in terms of the real world. A metacognitive processor is the stand in for Graphic Processing Unit. Why? Because it is specialized processor for graphics applications, which in our real world turns out to exactly what A. I. needed. A. I. as a graph...
 
Last edited:
Okay, why not transistors?

This has been extremely bugging me. Just figured out the answer. Why relays instead of transistors?

Relays are mechanical, not electronic. In other words, they move. It just isn't a question of being multistate, but of Faster Than Light computing power... which means a vacuum, and a pretty total vacuum.
 
When it came to designing the Enterprise D for TNG, Roddenberry was very conscious of the fact that its computer should be significantly more advanced that the Kirk-era ship (to show that nearly a 100 years has passed), not to mention significantly more advanced than real world 1980s computers.

He assumed that by the 24th century a ship's computer would be so advance that it would automate nearly all of the ship's functions and require little input from humans. When human input was needed, he pictured that being achieved entirely by voice. (In 2026 we already have people speaking to Siri and Alexa to control their devices, so this will probably happen sooner than the 24th century.) Accordingly, Roddenberry wanted the bridge having no user control panels whatsoever. Everything from steering the ship to scanning, launching probes, firing phasers, adjusting frequencies of the shields, etc. would be accomplished through voice commands.

He was talked out of this, largely because it was assumed viewers would want to see crew piloting the ship. He agreed, but insisted there be only two stations, con and ops, to show the ship was largely automated. (The aft stations were meant for ad hoc usage and not necessary for day to day operations.) He was adamant the bridge not be staffed by a large crew hunched over control stations, specifically because that would make its computer look less advanced.
 
A bridge (or even an office today) that has more than one person needing to work with a computer that only has a voice interface would quickly become noisy and confusing.

If almost everything was automated on a starship, you wouldn't need to give many voice commands. For example, if you tell it to modulate the shield frequency it would figure out all the details by itself. You wouldn't need to enter the calculations. You speak one sentence and sit back.

In the real world, the more technology advances, the more user friendly it becomes. In the old days it was a ten-step process to download and install software to your PC requiring the user to make all sorts of decisions. These days, just scan a QR code and your phone it does everything itself.
 
Last edited:
A bridge (or even an office today) that has more than one person needing to work with a computer that only has a voice interface would quickly become noisy and confusing.
Already solved.
M. I. T. The Oxygen Project.

Imagine a chip that could contain 1024 20 mgz Intel 80386 CPUs...with a variable wiring structure to optimize problem solving.

Open microphones that track each individual crew. And ignore via triangulation anything else. It knows where you are and what interactions are permitted.

Real world problem: it took back then too much power. A cell phone with a single one of these chips could do wonders by the standards of the time.
 
Solved for the computer maybe, but not for the humans. Still noisy for them if they're all talking to the computer trying to get stuff done.
 
Your voice is unique. Your coworkers voices are unique.

Multiple microphones can pull out your unique voice from hundreds.

Noise suppression techniques mean that more than a couple of feet away that you can't hear your coworkers.

Now produce a Tricorder...it pays attention only to what it is set for. Keeping in mind that the Tricorder sequences through predetermined settings.
 
The Enterprise largely being automated would explain why it's no big deal giving Wesley Crusher a bridge position.

On the topic of voice, we're shown that only Soong androids or changelings are able to voice-mimic officers to hijack ships, yet we already have AI youtube videos mimicking voices that I'd guess may confound a computer as well. It's possible the Federation has some sort of "magical" voice detection tech where it will always identify the speaker, regardless of how good the impostor is (again, unless you're a Soong android or a changeling) or how much the speaker tries to disguise the voice. In regards to the latter, Kirk says as much to Kodos, but it's possible he's bluffing since afterwards Kirk concludes that the computer couldn't lock a 100% voice match onto Kodos (even though it really was Kodos and he didn't even bother trying to disguise his voice because of his guilty conscience)
 
A bridge (or even an office today) that has more than one person needing to work with a computer that only has a voice interface would quickly become noisy and confusing.
Already solved.
M. I. T. The Oxygen Project.

I don't think this would be a problem, and I don't think you need to point to something as exotic as an MIT project to make the case.

Under Roddenberry's vision, there wouldn't be a lot of people talking. The whole point of a voice-controlled system is that it's largely automated. You'd only need a small bridge crew, and those few people would be able to control the ship with short commands.

"Computer, take us to Planet X at warp 3." ... "Computer, scan that anomaly and report." .. "Computer, evasive maneuvers." ... "Computer, modulate the shields to account for the Romulan disruptors." That's it. No need for you to enter calculations because the computer would figure it out.

So if you're picturing a bridge large crew constantly chatting at their computer screens, you're missing the point. Instead, imagine perhaps four people sitting on the bridge and rarely speaking to the computer. Logically this is plausible, but dramatically it would be boring.

Also, in 2026 voice interfaces can already distinguish between voices. If you have a smart device (Siri, Alexa, etc.) it knows to respond only to you. Presumably this tech would be even more accurate 300 years from now. One the Enterprise bridge, the con could be programmed to respond only to the helmsman's voice, tactical to the security officer's voice, and so on, with the captain's voice having override authority.
 
Okay, why not transistors?

This has been extremely bugging me. Just figured out the answer. Why relays instead of transistors?

Relays are mechanical, not electronic.
I think there may be a return to mechanical computers...we hear clicking in TOS....MEMs on microphones.

There is some talk about analog a.i. and ising machines:

My head canon says that there was a cyberpunk type of crisis, and that a hack proof mechanical system resulted....or something
 
Computation by…snap bracelet?

If you have ever seen the inside of an old automatic transmission…it even looks like a chip.

With the spinning contraption, we can now say any old reel-to-reel props don’t necessarily just have magnetic tape.
 
If you are not already a member then please register an account and join in the discussion!

Sign up / Register


Back
Top