Got an interesting hypothetical question here. Suppose I am anticipating contact with a potentially hostile and extremely intelligent extraterrestrial entity in a controlled scenario. I will be required to perform tasks that would be greatly aided by access to as modern a computer as possible.
However I would very much prefer to limit the entities knowledge of standard modern computer systems. Ideally I want to use a computer which has software and if possible hardware that is divergent as possible from the typical windows/apple/linux computer. Is there any available computer system that would fit the bill? and if not would it be possible to design a system that would work?
Possibly some of the US intelligence communities have high-performance computer systems that for security purposes are deliberately designed to be as incompatible with standard systems as possible?
Granted a programmer will probably be along to give a better answer than this, but my first thought would be, make a computer on an all integrated board and write and OS with the specific drivers for only that hardware hard coded in. Eliminate all support and plugins for any kind of plug and play or modular architecture, write software for it that only works on that hardware platform, keyed to that CPU/motherboard type. Use an exotic file system that requires additional check bits or encryption, something that normal file systems do not making them totally incompatible for a minor shift in efficency.
The concept of digital computing would probably not be lost on any spacefaring race and given time they could probably reverse engineer things and or learn how to write their own operating system with resources freely available on the internet. In some ways, once they can read our language, the internet will probably very easily let them know they are being fucked with and make them very unhappy.
So is the idea to have a computer that is powerful enough to be useful but different enough from mainstream stuff that if the aliens get hold of it they can’t use it to infiltrate current systems?
You need a Powerbook XXXX. That way when the inevitable hostilities occur you’ll be able to interface with the alien computer system and use a virus to disable their entire defence grid.
I’m not sure what existing system would be best, but it should be fairly straightforward to design a system that is almost arbitrarily divergent from standard computers. The primary reason that modern PC architecture is largely similar across platforms is because we want it to be compatible, not because there is some fundamental requirement that our PCs are designed a certain way.
I would use a thin client/terminal and configure it so that it can only communicate with a server at the home base. All applications and data are stored on the server, where all processing is done. The thin client exists solely to send user input back to the server and to show the user interface as received back from the server.
Now all you need is encryption and a non-standard communication protocol and the aliens will have nothing they’d find useful.
If speed isn’t vital: Make a computer out of an FPGA programmed using a custom core: An FPGA is a piece of hardware that can be physically reconfigured using software; a ‘core’, in this context, is the software that turns the FPGA into a specific kind of chip, which would be a CPU in this case; and a ‘custom core’ is a core that doesn’t work like anything being mass-produced. The great thing about an FPGA is that it’s possible to use one to create a chip that’s one-of-a-kind: If nobody reuses that core, the software that runs on that chip will never run on anything else unless they learn how to write a piece of software that emulates that core and works on other hardware.
This will unavoidably be slower than a modern Pentium CPU. However, it may well be fast enough to work, and it will be more-or-less useless to those evil aliens trying to learn details about how most modern human hardware works. It’s the absolute best solution from a secrecy perspective.
Of course, the software should be as stripped-down and as custom as possible. I’d use a simple hypervisor design with applications designed to run directly on that, so you give yourself the ability to multitask without giving the aliens much of an idea how our OSes are built.
This isn’t an entirely new idea, however: IBM called the basic idea of applications running directly on the hypervisor ‘service VMs’ back when they invented them in the 1960s. They’ve never been popular, though, and they have the advantage of not having most of the complexity of a full OS, and the software is tied more-or-less utterly to the hardware.
Sounds like a modern System Z Mainframe would work. Quite different, but already existing & working, much faster & more powerful, and a much more secure system.
I think most recent ‘supercomputers’ have custom hardware and software, not directly compatible with anything outside that company’s products. So I think what you want is the best Cray can provide, with a custom GUI front end.
What about a smart phone? Its not clear from the OP exactly what you need to keep secret or exactly what tasks you need to do, but a modern smart phone should be sufficient to access the internet for information. Swallow the SIM if you feel you need to keep internet access away from the aliens, then the phone is nothing more then a calculator, music player and electronic address book.
I think that’s a good idea. You might want to consider making sure the phone you bring is locked and jailed (opposite of unlocked and jailbroken), making it more than trivial for the aliens to interface with.
A computer is essentially a logic processor, basically a physics-powered flowchart. They differ in their processing speed and user interface, but at its core a computer is a computer. Any hyperintelligent species capable of understanding math, logic, and following instructions can make or use a computer. You can’t hold that back from them any more than you can prevent them from learning that “1+1=2” or “C is true if and only if A and B are both true”. As far as we know, the underlying laws of physics and math aren’t biased towards any particular species… but the ease of use of any given device certainly can be. Given that, I would tackle this not from a computer-science perspective but from a biological or perhaps cultural one. You’re basically designing a test for humanness.
You want an interface that’s as human-friendly and alien-unfriendly as possible. What separates humans from your hypothetical aliens? Do they lack fine motor dexterity, for example? If so, a keyboard-operated computer would be very difficult for them to use. Are they colorblind or do they see in different wavelengths? If so, a human-oriented display could appear blank (or contain false information) to their eyes. Can your computer simply have a fingerprint or DNA sensor to only allow humans? Do the aliens experience different emotions and brain states? If so, how about a real-time mind-mapping device that only allows non-coerced human access?
Otherwise, human cultural things could be completely incomprehensible to an alien without prior experience. Could they even read English, for example? Can you put up culture-relevant passwords? (Hum a national anthem, identify this rock musician, etc.)
I was thinking of something similar. Powerful, but not very mainstream. Another possibility is an Itanic (Itanium officially) based computer. Reasonably fast, and non-standard, and if the aliens try to copy it they’ll go down an expensive blind alley.
If you are going to customize things you can put together a microprogrammable machine and give it a very weird instruction set.
Another problem is that you don’t want them copying your process technology. You might want to build something in a very antiquated fab, which supplies processors for toys and greeting cards, or you can build something in a fab at Sandia which makes rad hardened stuff with old but very reliable technology.
FPGAs are good for disguising the architecture, but maybe not so hot for disguising process technology.
For software, I bet OS 260 would still run on the Z-series, or maybe you can microprogram the 360 instruction set (IBM did for lower end machines) and run stuff on there. There are some old IBM OS’s that would set the aliens back generations! Or, give them Windows 2.
100 monkeys with 100 TRS80 pcs. It’s not likely to work, but the aliens will be so amused by the monkeys, they won’t notice you doing all the work on your cell phone.
The problem is the ‘floppy disk fallacy’: Just because you don’t use IBM mainframes doesn’t mean nobody important does. In fact, System z mainframes are used in a lot of really important applications, so they’re not the kind of tech you’d want hostile aliens to have.
This is similar to my solution, above. My solution might be improved by a touch of this, in fact, but you hit the fact that FPGAs are slow to begin with and adding a layer of microcode won’t help that one bit.
This might not play nice with the FPGA concept, but if you have the resources to hire a fab to make one chip (or perhaps a few chips, to account for defects and testing) it’s a good way to go: Bizarre/obsolete process, Turing tarpit ISA, and heavy microcode, all wrapped up with a stripped-down OS and controlled with an abstruse command language.
(My god, don’t tell me… IBM designed the AS/400 for evil aliens!)
Do you mean OS/360? Because I’m fairly certain it would. IBM takes backwards-compatibility even more seriously than the Intel x86 people.
Or, in the case of VM, move them to a spot just ahead of the curve.
