# Computer control of a machine for exploring Mars

> An annotated edition of this paper on Fermat's Library, where scientists and readers explain its ideas, equations and history in the margins. Annotated edition: https://fermatslibrary.com/s/computer-control-of-a-machine-for-exploring-mars

- **Authors:** John McCarthy
- **Published:** 1964
- **Topic:** Computer Science
- **Original:** https://s3-us-west-1.amazonaws.com/fermatslibrary/uploads%2F1772425580101-nm8ug935h39-b37da627834b5f258c5e1edc8b7449e2%2Fmars_computer.pdf
- **Pages:** 7
- **Annotations:** 12 comments in 12 threads
- **Featured:** Fermat's Library Journal Club
- **Annotated edition:** https://fermatslibrary.com/s/computer-control-of-a-machine-for-exploring-mars

## Annotations

### Annotation 1 (page 1)

Beside the passage: “i COMPUTER CONTROL OF A MACHINE FOR EXPLORING MARS”

**Annotation** by João Batalha (2026-03-03):

> ### Autonomous Space craft in the 1960s
> In June 1964, when John McCarthy wrote this memo, no autonomous vehicle had ever operated on another celestial body. All prior space probes - such as the Soviet Luna 1–3 flybys that photographed the Moon’s far side in 1959 or NASA’s Mariner 2, which flew past Venus in 1962 - were non-mobile craft equipped only with fixed, hard-wired instruments. These probes collected data during a single pass or crash landing and transmitted it back in real time or via simple recorders; they had no on-board programmability, no mobility, and no ability to navigate or adapt to terrain. Soft landings on any extraterrestrial surface had not yet been achieved anywhere (the first would be Luna 9 on the Moon in February 1966), and the concept of a rover was absent from all mission designs.

### Annotation 2 (page 1)

Beside the passage: “by John McCarthy Computer Science Division Stanford University”

**Annotation** by João Batalha (2026-03-03):

> **John McCarthy (1927–2011)** was one of the founders of artificial intelligence. He coined the term *“artificial intelligence”* in the mid-1950s and helped launch the Dartmouth summer project that effectively started the field. He later created **Lisp** (late 1950s).
>
> He spent part of his early career at **MIT**, then went on to build Stanford’s AI effort, and from 1965–1980 he directed what became the **Stanford Artificial Intelligence Laboratory (SAIL)**.
>
> McCarthy received the **Turing Award** in **1971** for foundational contributions to the field. This **1964** memo reflects a career-long ambition to formalize common sense and logical reasoning within machines, aimed at solving complex real-world problems like autonomous planetary exploration.

### Annotation 3 (page 1)

Beside the passage: “1. Advantages of a computer controlled system.”

**Annotation** by João Batalha (2026-03-03):

> "computer controlled" as opposed to fixed, hardwired sequencing and simple relay logic.

### Annotation 4 (page 1)

Beside the passage: “3. What we can feasibly program the machine to do given the present state of work on artificial intelligence.”

**Annotation** by João Batalha (2026-03-03):

> John McCarthy had coined the term “Artificial Intelligence” just 9 years earlier (1955) when proposing the Dartmouth AI project.

### Annotation 5 (page 2)

Beside the passage: “In 1969 or 1971 we can land 5000 pounds on Mars. The machine we land can have 300 watts of power, can communicate from 10,000 to 100,000 bits per second back to earth according to the distance between earth and Mars. The machine shoul,.l be able to operate for a year. These facts ',Jere taken from a Gp.n:,~ral …”

**Annotation** by João Batalha (2026-03-03):

> In 1963 NASA commissioned General Electric to study a Mars lander as part of the early Voyager program, using Saturn V rockets for a 1969 or 1971 launch. Codenamed “Beagle” (after Darwin’s ship), the study described a stationary 5,000-pound lander with 300 watts of power and a one-year lifetime. Its goal was a search for life, but it included no mobility and no on-board computer control. The projected cost exceeded $1 billion in 1964 dollars. McCarthy uses it as a technical baseline to argue that such a massive payload could and should be mobile.
>
> The ambitious Voyager Mars program was cancelled in 1967 due to budget cuts and Apollo priorities. It was later revived and scaled down as the Viking program, which successfully landed two stationary spacecraft on Mars in 1976 - the first soft landings there - carrying life-detection experiments but without mobility or the programmable computer control McCarthy envisioned.

### Annotation 6 (page 2)

Beside the passage: “1. To carry ou-'- as thorougi1 a search for life as possible, i.e., to maximiz.e the probability that if life exists on Mars, Beagle will find it.”

**Annotation** by João Batalha (2026-03-03):

> ### Public interest in Martian life
>
> When McCarthy wrote this memo, the notion that Mars might be home to intelligent life—or even thriving “colonies” of canal-building Martians—had only recently faded from mainstream belief. Widely reported newspaper headlines (e.g., New York Times, 1911: “Martians Build Two Immense Canals in Two Years”) and popular culture had fueled decades of public excitement about Martian canals and civilizations. The idea lingered into the early 1960s, even as most astronomers grew skeptical. Just thirteen months later, NASA’s Mariner 4 flyby (July 1965) sent back the first close-up photos, revealing a cratered, dry world and ending the canal myth forever.
>
> ![](https://i.imgur.com/r1KvUd5.jpeg) *New York Times cover from 1911 announcing the construction of canals on Mars*

### Annotation 7 (page 2)

Beside the passage: “2. Many of the experiments (~an use common facilities of manipulation, picture recognition, etc.”

**Annotation** by João Batalha (2026-03-03):

> McCarthy is pointing out that many “experiments” are really applications on top of the same core capabilities: sensing, manipulation, and vision. That framing mirrors early time-sharing and OS design: build shared system services (I/O, scheduling, device control) once, then let many programs reuse them, instead of hardwiring each experiment as a one-off.

### Annotation 8 (page 3)

Beside the passage: “30 If Beagle works for a year the results of the early experiments will make changes desirable in later ones~”

**Annotation** by João Batalha (2026-03-03):

> McCarthy observes that if the Beagle operates for a full year, early experimental data will naturally suggest changes to later observations. This advantage is possible only with a stored-program computer that can be reprogrammed from Earth.
>
> This foresight was fully realized in the Voyager missions launched in 1977. The spacecraft flew pre-set trajectories. Yet ground controllers uploaded thousands of new command sequences after each flyby. Discoveries such as Io’s active volcanoes and Jupiter’s faint ring system led to completely revised imaging plans for the later encounters. Different filters, exposure times, pointing directions, and entirely new targets at Saturn, Uranus, and Neptune were used. For instance, close-up pictures of Neptune were never part of the original mission plan. New logic was uploaded to rotate the entire spacecraft slowly during exposures so that long-exposure images of the planet and its rings would remain sharp despite the spacecraft’s motion. The result was far richer scientific returns than any fixed pre-launch program could have achieved.
> ![](https://assets.science.nasa.gov/content/dam/science/psd/photojournal/pia/pia02/pia02245/PIA02245.jpg) *Close picture of Neptune - only made possible by software uploaded to the Voyager probe after launch*

### Annotation 9 (page 3)

Beside the passage: “1.1 weight - 100 lbs. 1.2 volume - 2 cu. feet. 1.3 memory cycle 1 ~ sec - add instruction 2 ~ sec - floating multiply - 10 ~ sec. 1.4 power consumption - 40 watts. 1.5 memory 130,144 - 48 bit words.”

**Annotation** by João Batalha (2026-03-03):

> In June 1964 these targets were ambitious yet within the realm of active research. Ground-based scientific computers like the IBM 7090 offered a 2.18 µs memory cycle, up to 32 768 words of 36-bit core memory, and add times in the same range as McCarthy’s 2 µs, but they filled entire rooms and consumed tens of kilowatts. No spacecraft had yet flown a stored-program computer at all. The Apollo Guidance Computer, then in early design, would reach flight with an 11.7 µs cycle time, only 2 048 words of erasable memory plus 36 864 words of fixed ROM, and 55 W power consumption - already considered a major engineering triumph.
>
> McCarthy’s goals of a 1 µs cycle, 2 µs add, 10 µs floating multiply, 40 W total power, and 130 144 words of 48-bit memory (roughly 780 KB) were therefore pushing the limits of what seemed feasible for a radiation-hardened, one-year Mars mission. In practice they proved challenging to meet for the era. The Viking landers that finally reached Mars in 1976 flew dual Honeywell 24-bit computers with only 18 KB of plated-wire memory each, smaller and slower than McCarthy had hoped. The gap between laboratory performance and space-qualified hardware would narrow over the following decades.

### Annotation 10 (page 3)

Beside the passage: “2. If possible, the system should nat use mechanical secondary storage, e.g. tapes or drums. They make reliability difficult.”

**Annotation** by João Batalha (2026-03-03):

> McCarthy recommends that the Beagle computer avoid mechanical secondary storage such as magnetic tapes or drums. Moving parts are vulnerable to the extreme vibration, temperature swings, and radiation of a Mars mission and would sharply reduce overall system reliability.
>
> In practice, several early spacecraft did fly magnetic tape recorders despite this risk. The Gemini spacecraft (1965–1966) used a tape drive to load programs into core memory. Mariner probes, the Viking landers (1976), and the Voyager spacecraft (1977) all carried digital tape recorders to store camera images and science data for later playback when Earth was not visible. These devices worked but occasionally caused problems (including mechanical wear on Voyager’s recorders decades later). True drum memory, however, was never used in any flown spacecraft computer. McCarthy’s caution proved well-founded once solid-state memory became practical in later missions.

### Annotation 11 (page 3)

Beside the passage: “3. The system must be able to recover from programming errors in programs that carry out particular experiments. Otherwise, it will be impossible to allow the wide variety of programs necessary to make use of the flexibility of a computer based system. In particular, it would be difficult to allow the revision of …”

**Annotation** by João Batalha (2026-03-03):

> McCarthy requires that the Beagle computer survive errors in individual experiment programs without bricking the whole machine. Without this protection, the flexibility of running many different programs and uploading revisions from Earth after early results arrive would be too dangerous.
>
> He draws directly on techniques then emerging in time-sharing operating systems. These systems introduced mechanisms such as user mode versus executive mode, memory protection, and timed interrupts to isolate user programs and prevent one faulty program from crashing the whole computer. In 1964 such safeguards were brand new; most computers still ran unprotected batch jobs where a single error could stop everything. McCarthy, who had proposed time-sharing concepts as early as 1959, saw that the same mechanisms could enable safe, adaptive programming on a distant spacecraft. This insight was far from obvious at the time.

### Annotation 12 (page 4)

Beside the passage: “10 The time-sharing executive - divides the time among the application programs.”

**Annotation** by João Batalha (2026-03-03):

> McCarthy’s proposed software structure is a remarkably accurate blueprint for the modular operating-system architecture that would dominate computer science for decades. By separating the core “platform logic” of the machine from the specific “tasks” it performs, it anticipates several foundational principles of modern systems.
>
> **Kernel and scheduler:** the “time-sharing executive” acts like an OS kernel scheduler, allocating CPU time across programs and ensuring critical system work (like thermal control and comms) keeps running even if an experiment program misbehaves.
>
> **Hardware abstraction and drivers:** putting device operations into shared subroutines with built-in safety checks is an early version of device drivers and a hardware abstraction layer. The goal is to protect motors, sensors, and actuators from being damaged by buggy experiment code.
>
> **Separation of concerns:** splitting “housekeeping” (system services) from “experiments” (applications) creates a clean boundary so domain experts can write specialized programs without also managing low-level realities like power budgets, telemetry, or mobility control.

## Full text

Extracted from the PDF. [Annotation N] marks where each annotation above sits.

#### Page 1

STANFORD ARTIFICIAL INTELLIGEkCE PROJECT Memo No. 14

June 15, 1964

i COMPUTER [Annotation 1] CONTROL OF A MACHINE FOR EXPLORING MARS

by John McCarthy [Annotation 2] Computer Science Division Stanford University

Abstract: Landing a 5000 pound package on Mars that would spend a year looking for life and making other measurem~nts has been proposed. We be+ieve that this machine should be a stored program computer with sense and motor organs, and that the machine should be mobile. We discuss the following points:

1. Advantages [Annotation 3] of a computer controlled system.

2. What the computer should be like.

3. What we can feasibly program the machine to do given the present state [Annotation 4] of work on artificial intelligence.

4. A plan for carrying out research in computer controlled experiments that will make the Mars machine as effective as possible.

The preparation of this memo has benefited from discussion with E .• Fredkin, J. Lederberg and M. L. Minsky.

The research reported here was supported in part by the Advanced Research Projects Agency of the Office of the Secretary of Defense (SD-183)

#### Page 2

COMPUTER CONTROL OF A MACHINE FOR EXPLORING MARS

by John McCarthy Comp~ter Science Division Stanfor.} IJni vers i ty

In 1969 or 1971 we can land 5000 pounds on Mars. The machine we land can have 300 watts of power, can communicate from 10,000 to 100,000 bits per second back to earth according to the distance between earth and Mars. The machine shoul,.l be able to operate for a year. These facts ',Jere taken from a Gp.n:,~ral Electric study called Beagle [Annotation 5] after the ship that took Darwin around the world. The Beagle study does not discuss the possibility of making the machine mobile, but we believe this can and should be done even if the power limitation makes it go very sluwly. We shall f'olltjw G. E. 's lead and call the machine the Beagle.

The Beagle should obl:.ain as much Llformation as it can about Mars and radio it back to earth. Naturally, the most interesting question is whether there is life on Mars and if so what it is like. Therefore, we can set forth three goals.

1. To carry ou-'- as thorougi1 a search for life as possible, i.e., to maximiz.e [Annotation 6] the probability that if life exists on Mars, Beagle will find it.

2. If life exists to find out as much as possible about its chemistry, physiology, ani ecologya Chemistry will be emphasized because the same means that de+.Iect life may also be Ilsed to study its chemistry.

3. To find out anything about the environment of Mars that will help future exploration, especially mannej expL)ration.

Why Computer Control ot the Beagle:

Up to now space r1·~.;bes have consisteci .)1' a collection of separate experiments sharing pr:='pulsion power supply and telemetering. We believe that Beagle will be much more effective if it is a computer with sense organs and mnt.c'r organs and the experiments are represented by ,~omputer programs €:t.,;b of which uses the sense and motor organs in a co-or1inated way. Beagle differs from previous space experiments in a nurriber of ways that :lre relevant to this preference.

1. A large numbe r of sense and motor ·.)rgans can 'be included in a 5000 pound machine.

2. Many of the experiments (~an use common facilities of manipulation, picture [Annotation 7] recognition, etc.

#### Page 3

30 If Beagle works for a year the results of the early experiments will [Annotation 8] make changes desirable in later ones~

These needs of the Beagle mission can best be met by a computer controlled system. A brief statement of the reasons follows:

1. The control circuitry of each sensory or motor device can be reduced to a minimum if the whole system is computer controlled.

2. The strategy of each experiment can be chosen freely by writing suitable programs even after the hardware decisions have been made.

3· New programs can be written and transmitted from the earth even after Beagle is on Mars.

The Computer and Its Programming:

In this section we shall discuss the features that the Beagle computer should have.

1. It should be light, compact, fast, have a large memory, and be reliable. We shall not discuss how these features can be achieved in this paper, but many companies are working on the problems involved, and we are quite sure a suitable computer will be available.* Suitable parameters might be

1.1 weight - 100 lbs. 1.2 volume - 2 cu. feet. 1.3 memory cycle 1 ~ sec - add instruction 2 ~ sec - floating multiply - 10 ~ sec. 1.4 power consumption - 40 [Annotation 9] watts. 1.5 memory 130,144 - 48 bit words.

If these goals are too hard to meet, some compromises are possible, but even higher performance might be helpful.

2. If possible, the system should nat use mechanical secondary storage, [Annotation 10] e.g. tapes or drums. They make reliability difficult.

3. The system must be able to recover from programming errors in programs that carry out particular experiments. Otherwise, it will be impossible to allow the wide variety of programs necessary to make use of the flexibility of a computer based system. In particular, it would be difficult to allow the revision of programs from the earth on the basis [Annotation 11] of preliminary experimental results if an error in such a revision could cripple the whole machine.

* I don't want to suggest that reliability will come automatically, only that I don't have anything important to say about it.

2

#### Page 4

The ability to recover from progranmling errors can be achieved by the same devices as one beginning to be used to make time-sharing monitor systems proof against user errors. The necessary features are available on the Digital Equipment PDP-l and PDP-6 computers, on the IBM 360 computer and partially on the IBM 7090 and 7030 computerso In fact, the Beagle computer should be operated with a time-sharing system, although the Beagle monitor must differ substantially from time-sharing systmes oriented towards computation centers. The important features of time-sharing systems are the following:

1. The system has a user mode and an executive mode. When in user mode the use of input-output instructions is inhibited and attempts to change memory outside an area reserved to a particular program leads to interrupts to the executive program.

2. A clock leads to an interrupt of the executive every so often anyway 0 (Say, every millisecond). The executive then decides what program sho'.lld be executed next for a quantum of time ..

3. Input or output devices generate interrupts to an appropriate part of the executive program wnenever input becomes available or an output device is ready for more.

The core of the executive prograrr~ niust be absolutely debugged, but protection can be provided against erl"'ors in large parts of the executive (eogo, the programs "that handle input-output devices) by allowing earth generated interrupts to a part of the executive that can be instructed to make changes in the rest of it.

We envisage the program to be divided into four parts.

10 The [Annotation 12] time-sharing executive - divides the time among the application programs.

2. Housekeeping programs. Handle communication with earth, temperature control management of the energy and supply, control of the motion of the machine~

3. Programs for operating deviceso Used as subroutines by the programs that run experiments. Normally contain checks to make sure the devices are not damaged.

4. Programs for running experiments. These are written under the supervision of the experts in the field in which the experiment is performed. The time-sharing system permits them to be wri.tten independently of each other.

Mobility:

'l'he effectiveness of the Beagle will be greatly enhanced by mobility. There are two difficultieso First, an average power of 250 watts will not 3

#### Page 5

move a 5000 lb. vehicle very fast, and not all the power is available for that purposeo Second, the motion cannot be directly controlled from the earth because the response delay varies from a little over six minutes to almost 25 minutes.

The first difficulty can be overcome by accepting very slow progress (e.g., 10 cm/sec to 100 cm/sec depending on terrain) at times when the experiments and information transmission require very little energy_ The second problem cust be solved by developing computer programs capable of steering the vehicle past obstacles over different terrains.

Mobility is important for the following reasons:

1. Beagle might land in an unsuitable place, e.g. on bare rock or in a ditch.

2. Beagle should be able to look for high points from which to transmit pictures of the landscapeo

3. Features that looked interesting in pictures could be examined at close range.

4. The search for life will be more effective if Beagle can go look for ito

Artificial Intelligence:

Research labelled artificial intelligence is aimed at making computers perform tasks that require intelligence when performed by humans. The exploration of Mars involves many such tasks. If the artificial intelligence problem were completely solved we could expect to send a computer to Mars with no control from earth and have it send back all the information that could be acquired by a large manned expedition. In fact, it is very unlikely that results comparable to manned exploration will be achievable by computer controlled machines within the next twenty years. However, many of the subsidiary tasks are within or near the present state of the programming art especially if the machine can be instructed from the earth if it gets stuck.

Some of the tasks are:

1. Controlling the telemetering so as to submit information at the maximum rate compatible with the orientation of Mars and the distance from Mars to earth. Information of lower priority can be saved for later transmission at times when high priority messages have to be sent.

2. Compression of information. Sending only deviations of an instrument reading from its expected value based on previous readings. Picture compression is more difficult but some results have been achieved.

30 Picture recognition. In various forms, picture recognition is required for a number of Beagle's tasks. Some of these are: 4

#### Page 6

;.1 Recognizing types of terrain and obstacles so that Beagle can obey orders to move.

;02 Recognizing the lands of materials it has been ordered to collect for analysis.

;.; Co-ordinating the devices that pick up samples and subject them to analysis.

4. Motor co-ordination, co-ordination of the "hands" and wheels and "legs".

5. Other experimental strategies.

The assumption that the computer can be programmed to achieve the above goals relies heavily on the ability to reprogram it from the earth when unexpected conditions are encountered. We do not expect the state of work in artificial intelligence by 1970 to make the following feasible.

1. To put in the program our concepts of what is interesting.

2. To define life well enough so that the machine would recognize any form of it.

;. To make the program adaptable to any terrain without further instruction. e.g., swamps or mountains.

Projects:

The problem of making good use of a 5000 pound payload is very difficult. A number cf investigations should be started right away if the Saturn V rocket is to be used when it becoffies available 0 Some of these projects are:

1. Design of a suitable computer. We assume that the work on small, light, reliable and fast computers with low power consumptions is proceeding. We are less confident that the computer compani€s will come up with system designs suitable for the sophisticated programming that would be required. The Stanford Computer Science Division would be interested in helping with the order code, input-out structure, and system program design for such a computer.

2. Artificial Intelligence. Anything that can be learned about how to make machines behave intelligently will eventually be of use in unmanned planetary exploration. The most critical problem for Beagle, however, is the visual pattern recognition necessary for selecting and picking U.p samples and for steering a vehicle past obstacles to a goal. The more we can achieve in general purpose manipulation the less it is necessary to rely on Rube Goldberg contraptions for raising antennas, picking up samples, righting the machine after landing, etc.

5

#### Page 7

•

3. A ~ample collector. The mechanical engineering of a device for picking up and breaking and crushing samples should be undertaken soon.

4. A vehicle. The low power that is likely to be available calls for a special vehicle design. For example, a crab that uses the same organs for mobility and for picking up things may be appropriate.

We believe that work aimed at a prototype Beagle that can be tried out on earth should be started as soon as possible. We are eager to help with this.

6

---

Source: [Fermat's Library](https://fermatslibrary.com/) — https://fermatslibrary.com/s/computer-control-of-a-machine-for-exploring-mars. More annotated papers: https://fermatslibrary.com/llms.txt
