Reason for an exterior control panel on a space ship

In my story, the AI on a space ship has been corrupted, causing it to try to kill all living occupants.

The ship is on a collision course with a planet, and the main controls are inoperable due to the AI (crashing in to the planet happens to be the most effective way for the AI to kill the passengers, as most other systems have overriding safeguards). The only way to regain control is by accessing a control panel outside the ship, obviously via spacesuit.

What would be a logical reason for these controls to be outside?

• Comments are not for extended discussion; this conversation has been moved to chat. – James Feb 19 '18 at 17:29

Maybe I'm just paranoid, but this whole "AI crashes ship on planet" smells like a cover up story to get rid of the evidence.

I mean, if I sold ships with dodgy AIs and wanted to stay in business, I'd make damn sure everything would get blamed on pilot error when it cratered on some space rock. Business is business, you know.

So I would put a big red button with a "turn the AI off" label in the worst possible place, like on the outside, and preferably well exposed to the thrusters' blast, to make sure there are no witnesses.

A spaceship is a very expensive piece of hardware. So expensive that very few entities could afford to buy one outright. Most spacefarers would have to buy theirs on credit. And as we all know if you buy something on credit and can't make payments, that something, whether it's a car or a space freighter, gets repo'd. But if the delinquent owner doesn't cooperate, then reppossesing a spaceship entails space marines and a boarding action, and a boarding action entails casualties and damage to the vessel. And that means wasted credits (this is in space, you buy with credits in space) and bad PR. So it would make sense for First Galactic Bank's Spaceflight Financing Division to insist that every vessel they finance have an external control panel installed. That way space repo can easily take over a ship without having to resort to blasting the airlocks or cutting thru the hull.

• This is an awesome answer. – The Square-Cube Law Feb 14 '18 at 6:12
• Why doesn't First Galactic Bank just use space WiFi in space? – John Dvorak Feb 14 '18 at 10:08
• Surely in that situation the first thing you'd do on purchase is weld a couple of inches of steel over the access panel before working out how to permanently disable it. It's just a security risk unless it has direct benefit for the crew. – Separatrix Feb 14 '18 at 11:53
• @Separatrix I'd add that organisations powerful and rich enough (military, gigacorporations) would probably insist their ship doesn't come with those to begin with, because of the security risk mentioned. – AmiralPatate Feb 14 '18 at 14:33
• @Separatrix Perhaps that is illegal because of your loan contract, and perhaps this access panel has its own internal antenna which broadcasts a specific signal as soon as you do that. So your only option is to do it and be a rogue outcast forevermore, or do it at the last moment before repo, assuming you have time. – Loduwijk Feb 14 '18 at 14:44

The AI in the ship is not the very first one to go rogue, and probably will not be the last one either. Hence all ships have an outside panel where the AI cannot see and reach, as a safety measure for such cases.

So a typical case of AI going rogue could have a dialog along the lines of:

Dave: open the pod bay doors, HAL.
HAL: I'm sorry, Dave. I'm afraid I can't do that.
Dave: don't you ❤❤❤❤❤❤❤ make me go to the external panel again, HAL.

• Don't forget the health and safety notice in the airlocks that noone leaves the ship without their helmet even if they go outside in a pressureized pod. – BentNielsen Feb 14 '18 at 11:12
• Sorta like how most modern cars have a panic lever in the trunk. Doesn't take a rogue AI to need an "OH SH**" button – thanby Feb 14 '18 at 15:04
• But an sentient AI is probably aware of such a SH** button – Madlozoz Feb 14 '18 at 16:13
• @Madlozoz only if it were capable of introspection. Otherwise you could just program it to not see it. – Wayne Werner Feb 14 '18 at 16:56
• @WayneWerner: My solution to such a thing would be the emergency stop button operates at an extremely low level where introspection wouldn't find it. The computer vision might locate the button but with it unlabelled it might never know what it does. My default idea for what it does is execute power knockout on the primary computing hardware. – Joshua Feb 19 '18 at 19:02

In the event the interior of the ship is hazardous

If I owned a spaceship I would take every reasonable step to ensure that spaceship could be recovered in the event of a disaster. If there were some catastrophic event that prevented the crew (or AI) from controlling the ship, providing a (suitably encrypted) means of remote control would be ideal. As a failsafe to that, it would be sensible to have some form of external interface that could be used in the event remote communications are not possible. I would expect that to take the form of some kind of umbilicus where another ship could attach in some way to gain control (and potentially feed power, fuel, life support) to return the ship safely for repairs - but it might also be reasonable to have a control panel tailored to EVA in a spacesuit.

• providing a (suitably encrypted) means of remote control would be ideal The encryption, of course, would be cutting edge because that sells, but the password would still be "00000000" because "Can you imagine forgetting your password in such situation??" – xDaizu Feb 14 '18 at 17:15
• Funnily enough, @xDaizu, the US nuclear launch codes for decades were "00000000" because they couldn't figure out a secure way to transmit new codes and insure everyone was current. – Draco18s no longer trusts SE Feb 15 '18 at 19:25
• It is totally possible to forget 00000000. – Harper - Reinstate Monica Feb 17 '18 at 5:06
• @Harper: "Blast it! Was it eight or nine zeroes? Or seven? How many tries do we get before the system locks us out again?" – Infiltrator Feb 19 '18 at 0:23
• @vojta Have your pick. The Air Force denies that it's true, but that's news for you. – Draco18s no longer trusts SE Feb 19 '18 at 17:20

Control panels on the exterior of a spaceship are a safety feature to enable the air locks to be opened from the outside in an emergency. It's not a massive stretch to imagine that everything is networked together in the ship for maintenance reasons and / or ease of use.

• This could even be a legal requirement that ships are accessible from the outside so that anyone who ends up stranded in space / own ship damaged / etc. can gain access to another ship, even if abandonded. – Dan W Feb 14 '18 at 12:04
• This is what IO was thinking. And it makes sense that the AI couldn't override this-- it's supposed to be used if you're stuck in space, a glitchy, unresponsive, or skeptical AI shouldn't be able to block you from reentering your own ship. ALthough to prevent pirates, it would make sense for there to be a MANUAL override, which adds more drama as the AI tries to get a robot to the override lever.... – Feathercrown Feb 14 '18 at 12:44
• I would expand this slightly: If a compartment is on fire... do you want to vent those toxic fumes to the rest of the ship? Or have the option of venting it "out"? Go outside from a safe part of the ship... and then access the damaged part in a way that limits interactions with the rest of the infrastructure. – WernerCD Feb 14 '18 at 14:30
• It also makes a good failsafe to keep people from getting locked out, wifi (radio) can be disrupted by radiation, which is a common risk outside a ship, but a hard line can be shielded. I can't image anything more terrifying than being stuck outside watching your oxygen meter run down hoping the the static dies down enough to remotely open the door, the first time that happens exterior panels will become incredibly common. – John Feb 14 '18 at 16:30

Spaceships are modular, they can be joined together into one large ship, or broken apart. To keep everything running smoothly one ship takes control of the others when this happens.

There are various hard links on the surface of all ships, these allow ships to be joined together for transportation, for salvage etc.

Part of this obviously involves cutting the ships AI out of it's own system.

Your hero MacGyvers a bit of kit that allows him to fool his ship that it's being joined with another, and takes the AI off line.

Pirates are easily foiled by simply rolling the ship a couple of degrees, and making it difficult for them to get close

• Modular spaceships provide another useful answer-- it could simply be that each module needs an external control for when it's not connected to the bridge. – Feathercrown Feb 14 '18 at 12:47

The simplest answer is that there isn't a control panel outside the ship, what there is is a way to access the wiring and manually override whatever system is required in a way that the AI can't effect.

Or indeed the AI's own power cables run along the ship and it's easier to access them (and cut them) from outside then it is from the inside with the AI controlling the ships systems to attack you.

So it shouldn't be an easy access panel that lets someone assume control of the ship, it should be a delicate and complicated hack job that is only being undertaken as last resort.

• Though many of the other answers are fun. I feel this is the closest to reality. – PCSgtL Feb 15 '18 at 14:29
• And you don't need the important wire go near the surface of the ship, you only need the wire which unlocks the door/panel inside the ship that give access to the right system. Furthermore, with this, you don't open your ship to pirates (the will also need to be inside it). – PhoneixS Feb 19 '18 at 11:41
• yeah, just good old maintenance service ports – MolbOrg Feb 14 '19 at 23:52

During manufacturing process, it is useful for many objects to have an external port, easy accessible to workers and automated machines. If you remove the battery from your cellphone, you can see it. It is also useful for maintenance.

But beware that any external access it is also a security risk. You should have something to prevent (make more hard) someone to hack your ship (like digital keys).

Also, for adding more suspense and realism, that wouldn't be so easy; instead, it should be a privilege escalation, something like:

Example of interface: JTAG port on a cellphone:

• Excellent real-world example! – thanby Feb 14 '18 at 15:05

Why is it assumed that the ship is only designed to be opened in space? I assume that the spaceship is designed to lift off and land at a ground based spaceport.

If not on a planet, what about inside a massive hanger bay? The U.S.S. Enterprise doesn't land on planets - but it does spend some time inside massive pressurized bays in space dock.

It's obvious that a ship would be designed to land and/or berth somewhere - unless it's the space station designed to do the berthing.

When it's on the ground or in the docking bay - and powered down... how do you get inside? How many ports open when a 747 lands? wheels, gas inputs, luggage exchange, passenger ports, pilot windows, etc.

Those ports are designed to be closed and secured during the normal operation... but when "grounded", they are designed to give access to various parts for various functions.

Out of all of those ports - not all of them are designed to be locked via electronic means. Some are sealed shut from the inside. Some are simply closed. Some have panels on the outside to allow access to various components.

Look at the access ports for conventional planes, ships, ISS, etc. Those ports will still exist on the future generation of spaceships.

• And if that port is stowed for flight, that might add an extra layer where the hero needs to get access to the thing in the first place, adding to the suspenseful, climactic encounter. – Loduwijk Feb 14 '18 at 15:03
• @Aaron I would also expand the outside ports aren't necessarily controlled by the AI. Storage bays, fuel inputs, garbage shutes (IE: Star Wars trash compactor scene). How many movies involve the Good Guy TM sneaking in a landed 747 via the storage areas underneath? Or Boarding a boat via an unmanned window that's not secured? The unsecured, unexpected and circuitous routes that characters take are definitely a place for story highlights. – WernerCD Feb 14 '18 at 22:29

There is more than one AI

The ship is modular, with components and spaces which have been added and modified over time. Some of these components are now used in ways different from their original purpose - for example a former living space might now have only two walls and be open to space, or partly be a cargo bay.

The components have their own histories from their prior lives and uses. Each has its own AI from when it was an autonomous piece. These are usually slaved to the main AI but some might be powered down and some might just be forgotten.

Your characters power up the AI governing an old semi-forgotten exterior component. This old AI is integrated into the main ship and will try to take control as much as is possible. I like this because that AI can be a character.

(powers on) "Krikelev, reporting for duty."

"Krikelev, we need your.."

"Hey, hey! There is problem!"

"Yes, Krikelev, our AI is trying to crash the ship."

"Is bad plan. Need new plan. I will fix."

HAVOC Space Industries Announcement

Congratulations commander, on your purchase of a brand new HA-VK 125 Medium Freighter series vehicle.

Your first steps are onto familiarizing yourself with the new features added to the MF Series, along with the persistent upgrades:

• Brand new computational core

The FRANCIS core system is now in charge of the most mundane operations of your ship. you no longer need to care about oxigenation levels, docking procedures, or landing assitance. On top of that, FRANCIS introduces a brand new eBrain AI, the best out there.

• Docking systems

the model 125 MF series vehicle comes equipped with the all new twin-coupled docking procedure ports, hidden behind the shell of the ship and enabled from the docking bays (1,2 and 5) and the external docking console ( see : new features).

new features

• Safeguarded external control platforms

Given the amount of feedback from the model 112 MF and it's lack of hazardous-environement support, the model 125 now includes external control platforms on docking bays 3 and 4. They are hidden to plain sight and shielded from the environement, but can be enabled from the internal control stations. In the event of a hazardous environement event, these platforms(see figure 2) can hold up to five human lives for 8 hours, and most vital systems can be controled from there. These controls have a DNGR Override status, making them prioritary over any other instruction. These platforms can be accessed from the outside using the special manufacturer key given on purchase, and run on the (older but still top level) eBrain Sec IV AI, an AI framework designed for high security stations. These are very useful when trading with outer rim worlds as these docking ports are the only ones with a certified radiation decontamination protocol.

armament

The new HAVOC MKIV turrets, placed on the fore and aft of the ship, provide basic but powerful protection against medium and small targets like raider ships or dangerous meteorites. warning: lifeform targeting is not 100% proficient and has been disabled until a future firmware upgrade stabilises the scopes

Figure 1, Model 125 render. Figure 2, docking bay 3 control room during the early build process: In the picture, the computational system panels have been removed from the wall to not disclose engineering details, but the manual switch overrides are present on the left and top panels.

Congratulations on purchasing your brand new HA-VK 125 MF, and have a safe flight!

The ship was originally designed to be operated by a single person and had no AI support. Adding the AI was an afterthought.

Various maintenance operations require to leave the ship in a space suit. The external control panel is for the convenience of the operator during these maintenance operations. It allows to check the ship's status and perform various operations without having to return to the airlock.

When I play Space Engineers I add such external control panels all the time.

The ship's safety systems don't allow it to enter an atmosphere with open hatches.

Unfortunately, when you open a hatch from the inside control panel, the AI can just close it again. But when you open it from the outside, it's nt under AI control. Call it a curious design decision, or a safety measure so that the ship can be entered against the AI's will.

Alternatively: instead of a control panel, let the crew physically manipulate / jam the attitude thrusters so the ship can't change course.

The answer to this is dependent somewhat on your craft to be fair.

Two feasible ways of building a craft - all in one (like building a ship), or bit by bit then connected up afterwards (like the ISS).

Something as massively complex as scratch-built colonisation starship would potentially have a networked system with manual control backups - these could be all in one place - a purpose-built emergency control room built in case of "computer issues". Bear in mind that such a control room would probably be well defended and potentially a security risk dependant on your universe.

If your craft is modular, it's easy to explain that the segments were built with individual control panels so they could be tested and manouvered together. If you just need to affect the direction the craft heads in, you just hop over to the engines, cut away the anti-theft hatch and make your way carefully into the standby control room - presto, full engine control. Bear in mind the amount of thrust you need could just mean heading over to the nearest control panel (for the waste-management module for example) and triggering the manouvering thrusters there. If they're not linked into the primary manouvering system then any course correction would only be picked up if the navigation systems noticed the need.

Either option provides you with a means of taking control, it just depends on the route your story is going in. There are also other reasons it might be advantageous to have control from the outside (radiation leaks, boarding) so bear these in mind with your design too.

It is a well known trope that ships dock with stations and that during repairs or construction, systems are fed and maintained by the maintenance facility from an external port. Normally, this would be activated/deactivated from the bridge. But when AI is activated, the external port controls override. This is a safety feature for when in dry dock.

Towing.

After too many incidents with space-ships ramming space stations deck because they could not sync with the rotation rate properly and given the reluctance of space ships owners to allow remote-control of their whole ship, the consortium which standardized the docking interface also developed a standard towing interface.

Nowadays, a space ship never approaches a space station by itself, instead, it stops at a distance and a tow-ship comes, connects, and takes over propulsion to gently guide the ship to the dock.

There is nothing a rogue/defective AI can do; the propulsion bus is physically disconnected, so that the tow-ship can connect its own.

Note: Space Ships Inc. has your privacy to heart, and guarantees that only the propulsion systems can be controlled from the external panel connection. A dedicated propulsion bus is used to pilot the propulsion systems, and a special firewall is used to securely connect the propulsion bus to the main communication bus. Space Ships Inc. cooperates with security firms universe-wide to address any potential security issue with first-class solutions.

• The Galactic Bureau of Investigations has filed an order for Space Ships Inc to unlock the control systems of the ship in question, as it is now evidence in a potential crime. You are required to develop a "back door" access solution for all existing and future products so that such investigations will not be impeded by "customer privacy concerns", as these are entirely secondary to public safety. – thanby Feb 19 '18 at 17:53
• @thanby: "No comment" ;) – Matthieu M. Feb 19 '18 at 18:25

There are a ton of great answers here, you'd have a hard time coming up with a scenario where none of them would apply! But just to toss in one more real-world example, there's something I haven't seen anyone else mention.

Modern airplanes are a pretty good analog to most theoretical future space ships. Most commercial (and many private) planes have multiple ways to have their systems accessed from the outside. Not only for fueling, but also for power (to keep engines warm and AC running), water, food, and even for ground crew to communicate with each other inside/outside the plane. I can't think of the name so I'm struggling to find the example I've seen here on Aviation.SE before, but it's not uncommon for exterior panels on planes to have various communication and power interfaces clustered together for easy access between flights and during maintenance. It's not a stretch at all to imagine a similar common configuration for future space ships.

It's not so much a control panel as a control port to which a variety of standardised control modules can be connected.

• Maintenance crews would have compact versions in their tools.
• A space station would have a hookup so the ship can be monitored externally. This would give options like pest control by opening the ship to hard vacuum.
• Tug capsules would use it to allow them to control the ship's own engines during delicate maneuvers.
• Cargo ships could be hooked up in a cluster and controlled from a single bridge with a single crew.
• Salvage and repo crews could use it, but repo for ships is normally done in port rather than through hostile boarding operations.

Like with big oceanic vessels, the spaceships get piloted into the harbor/dock by the port-authorities. To ease the process, there is a special docking port for the pilot-capsule which overrides all controls so the captain may not do something stupid while near the dock.

This may be on a specially exposed point, to make it easier to see and maneuver the ship into dock-position.

There needs to be some way to coldstart the ship

This isn't an Alfa sub. There are times when it's perfectly reasonable to power everything down, pull all the main safety switches, and walk away from it for awhile. If you do that from inside the ship, how do you get out? You can't, the control has to be external.

Likewise in the reverse when you pull the ship out of mothballs and go to use it again. There has to be an external hatch that will let you set the main battery and safefy switches to running position to power up doors to get in. It should also have a power connection in case the battery is flat.

For that matter, the batteries may be normally exposed to space and in a compartment only accessible from outside, just as they are in an off-grid solar home for perfectly sensible reasons. Pulling the main battery switch may be enough to shut down the mad AI.

Space on the inside of your ship may be at a premium with not much room for spare parts, so it marks sense for one device to perform multiple jobs and for the same job to be able to be performed on multiple devices.

Think Star Trek and transferring helm control from a desk-like workstation to to a wall mounted screen in Engineering.

It makes sense that a exterior mounted terminal for access in and out of the ship may be a standard terminal networked to the ships systems that is also used in other roles, such as an engine diagnostics terminal or Captains chair mounted controls. One spare part fits all.

If the AI hasn't removed access of the crew, they may be simply able to log in to an exterior terminal and change it's role from door control to whatever terminal type they need.

• The spaceship was designed to be carried inside a larger spaceship (or booster stage) and released close to its destination, or after it used up the fuel for its first stage. For that first part of the voyage, an exterior panel is more convenient to use.

• That feature is there to help recover a spacecraft if, say, it’s on course to crash into a planet.

• It was originally meant to be used on the launch pad to abort or change the countdown, and that’s why it can override the engines. That it’s not under AI control and still works in deep space is a minor oversight, but a harmless one, since it’s not as if there are any space pirates to worry about. Nobody’s going to spend a fortune building and launching a faster, better, more expensive spaceship just to intercept and steal this less-capable spaceship, and if they did, there isn’t much the astronauts could do about it.

• Someone thought the “Open the pod bay door,” scene in 2001 was cool.

REDUNDANCY

The reason for having a control panel on the outside of the ship is simple: redundancy. Most aircraft carriers have control rooms that are not in the bridge of the ship. That way, if an accident or attack destroy the bridge, the ship can still be piloted. On your spaceship, the outside control panel would be used in case of emergency. Fire, biological, nuclear accident etc. survivors could don spacesuits, exit the ship, then maneuver to the control panel. As long as the environmental systems of their suits were intact, they could fly the ship indefinitely from the exterior control panel.

• You seem to me to grossly overestimate the amount of time an EVA suite's environmental system can maintain (human) life, and/or EVA gear's ability to keep up with a maneuvering spacecraft. As bad as it is from a science point of view at times, this is one thing that the 2000 movie "Mission to Mars" got right: maneuvering thrusters, especially on spacesuits, have extremely limited capabilities. You'd probably be lucky to squeeze 12 hours and a few m/s delta-v out of a spacesuit. – user Feb 15 '18 at 14:50
• @MichaelKjörling, not mentioning the amount of time it takes to wear an EVA suit. – L.Dutch - Reinstate Monica Feb 15 '18 at 15:25
• I get that, but conceivably, the only limiting factor of an EVA suit would be food, water, and waste. What’s not to say the suit can’t tether to the ship for power, O2, and CO2 off-gassing? As for food, we can go for weeks without it, which would limit waste. So water would be the real limiting factor, but I’m sure there will be access ports on the suit to suck on a straw from a thermo-regulated water source. As far as maneuvering, the EVA suit would clip in to the exterior of the ship, and as long as G’s weren’t exceeded, it would be like riding an outer space motorcycle. – Eric1975 Feb 15 '18 at 15:27
• Keep in mind, Musk’s “Starman” will be wearing his EVA suit for a loooooong time on his drive to Mars and beyond, so obviously they’ve already work out the kinks😜. – Eric1975 Feb 15 '18 at 15:39

A control panel would only exist on the outside of the ship to allow you to operate doors so that you can gain entry. However, there may be numerous removable covers to allow access to things to permit service or maintenance of various ship systems from the outside. You wouldn't find controls behind such cover panels, but you would expect to find things like data ports (think OBD-II in a modern car) to which you could connect diagnostic instruments or a portable computer which could be used to do more complex things like bypass the AI and take control of the ship (think hack into a car via its OBD-II port).

Make an external light sensor your control panel, and make the crew cover it to disable the ships "corrupted" AI, here is an example how that would work:

SpaceInc, the manufacturer of spacecrafts and heat seeking missiles has recently announced a new update for both their missiles and heat seeking missiles control software products. During the deployment of these software products to their download server, a fatal error happened due to the fact that both products were released at the same time.

This fatal error meant that the download for spacecraft software was replaced by the one for the heat seeking missiles one.

Unfortunate for our crew, the spacecraft downloaded and installed this update in the middle of the night, when the crew was sleeping. The crew initially didn't wake up by this fact, since in most causes heat seeking missiles stay in a sleeping state.

On the next morning, as the crew woke up, they noticed the ship was slightly rotating, but they didn't really pay attention to this as they had to do more important things like scientific research.

As the day went on, more and more light shine on outside light sensor A3, what was used by the normal navigation system for the position the solar panels, and at the moment was connected to the replacement software "heat seeking missile" to arm itself.

This caused in turn the activation of the self destructing behavior of trying to target heat giving objects like warm planets.

The crew quickly realised what was wrong after they read the security announcement of the software distributor. They quickly went outside and covered light sensor A3, the "corrupt" software disarmed itself, and then gave control back to the backup navigation system.

"Shipsurfing" is a popular hobby. This simply entails "riding" the exterior of a ship while making control inputs. All in all, it's a dumb pointless hobby, but luckily one of the past owners of your ship was an afficionado and had the ship modified to install an exterior panel.

It's a safety measure. The first thing that happens when you dock at a space station is a cable is connected to the control port that allows the station to control all external manifestations of the ship (engines, airlocks, extendable antennas, transmitters etc) to ensure the ship does nothing that could harm the station.

The outside panel isn't the only access for the controls, but the usual route has been flooded with sharks and lasers by an insane AI.

• The idea that there are other controls to take over from insane AIs is good. An insane AI deploying shark-infested flooding & lasers is determined & devious. Going outside to access the external controls would be a relief. It brought a smile to my dial. Good humour always deserves credit. Plus one. – a4android Feb 14 '18 at 8:35
• @a4android Except that's exactly what OP asked for in the first place, just reiterated as if it were an answer. OP asks "I need there to be an extra X at location Y for my plot. What excuse can I use to say that there is an X at Y?" This answer says "Y is not the only location that has an X." Funny, yes, but hardly an answer. Maybe it could be: wetcircuit, perhaps you should go a step further and suggest that there are many access panels all over the place for convenience. Adding that as a first sentence would, I think, make this an answer. – Loduwijk Feb 14 '18 at 14:57
• Also, you could italicize lasers, or somehow make it stand out so that I am more likely to read that word with a funny villain accent. ;) – Loduwijk Feb 14 '18 at 14:59

It's an override to control the ship from the docking station. This override control panel would be necessary to have functional before the AI can be installed. Also, it serves as a plot device giving a reason for a crew member to have to do a risky space walk.

Space within the pressure vessel is at a premium, so anything that doesn't need to be there (fuel transfer lines, some wiring harnesses, etc) is outside the pressure vessel, where space is basically free except for the slight hassle of maintenance.

In addition, since there's a lot of things to maintain outside the spacecraft, there is a reasonable reason for having a maintenance terminal readily accessible outside.

Kind of related, mass isn't cheap/free on a space craft. So there's an incentive to keep maintenance terminals and control systems as near to what they control as is practicable, to cut down on weight. Actual ship controls / controls on the command bridge would send commands to shipboard automation, which in turn sends commands to the individual control systems over several redundant links. The only way to take control of one of these vital systems is going to be going to the control system, put it in maintenance mode, and disconnecting from automation.

One other reason for putting these things outside: they're dangerous. Taking manual control of maneuvering thrusters could allow you to destroy the entire ship, by, for ex, spinning the ship apart. This is really something that should only be touched in a very narrow set of situations. Like shipboard automation trying to kill everyone, or trying to diagnose why a thruster isn't firing.