[iDC] The 50-Year Computer

[Patrick Lichty]

The 50-year Computer
Manifestos for Computational Sustainability, I

I have a proposition to make – when I am ready for my first mind/body 
transplant in 2058, at age 95, I want to be using the same computer I am 
today.  Upon first look, both may seem outlandish by today's standards, but 
I argue that the former is far less preposterous than the latter.  What I 
want, among other things, is a computer with a fifty-year life-span.  
However, attaining this goal will require a fundamental rethinking of our 
computational culture, and will likely be at odds with manufacturing agendas 
put in place in the mid 20th Century.  However, as we are confronted once 
again with issues of sustainability, paradigmatic changes are called for.  
The idea of a 50-year computer could be one of them.  My concerns that have 
led me to this proposition are based on theunnecessary planned obsolescence 
of computers, necessity vs. desire, and the huge amount of landfill that old 
computers contribute.

To elucidate on what I am proposing, I would like to draw some analogies 
analogies to the idea of the 50-year computer.  As a quick metaphor, 
consider the automobile.  If cars were operated under the same methods as 
computers, there would be three types of roads, each with its own kind of 
cars at this time (Windows, OSX, and Linux).  And, for at least two of these 
road types, every five or ten years or so, all cars made before may not 
drive on the current type of road.  For example, my father, aged 87 at the 
time of this writing, has a 1963 (one year after my birth) Studebaker Lark 
that he loves to shine up and drive every summer.  It does have some 
drawbacks, like the fact that is had to have seat-belts installed as an 
option, it lacks many of the amenities found on contemporary cars, and it 
occasionally needs new tires (which are often hard to find) because rubber 
gets hard with age.  It gets good gas mileage despite its eight cylinder 
engine, and drives quite well, if a little loose in the steering.  It is 
certainly not a “modern” vehicle by any standard, it is still basic, solid, 
fun transportation despite its age.  I didn't have to buy an upgrade to make 
it operable on current roads.  

People buy computers based on desire, not based on what they need. Many of 
my colleagues in academia, especially in history and the humanities, have 
only needed some Web capability beyond the basic office suite software in 
the last fifteen years, as is also the case of my father.  Keep in mind that 
I am talking about mass – market, general-purpose computing, and not more 
specialized media production applications, such as 3D animation, cinema 
editing, and the like.  Yes, making your own videos and putting them on 
YouTube is fun, but do we NEED it from a functional perspective?  I argue 
not.  This is not to say that consumer-level media production, high-
resolution games, and so on are not enjoyable, but this does not fit my 
proposition of an essential, appliance-grade, general computing platform 
that provides elements of “essential” computing for long periods of time.

In the 1980's when I was a field engineer for Tandy Computers (now defunct 
as a division, but was part of the Radio Shack consumer electronics chain), 
there was a LAPTOP called the Model 100.  It was entirely solid state, 
battery powered, had an extremely legible LCD screen, and was a known 
workhorse in the journalism field.  What did it do?  It had the basic office 
programs (word processor, database, spreadsheet) and a terminal program for 
online access, as well as a modem.  It also had a ROM slot for additional 
program chips. In my research, I have seen a number of them still in use 
nearly thirty years after manufacture, although obvious repairs such as 
keyboards (one keyswitch at a time, mind you) or displays have been done.   
The Tandy 100 is probably the first case of my knowledge of a basic platform 
for long-term general computation and an initial inspiration, if not just a 
fine metaphor for this computational model.

General Computing?
General computation refers to the creation of a standard for basic, 
ge3neric-use platforms with a set hardware criteria created to run the 
“road”, or operating system and its set of fundamental programs.  Since we 
are defining a standard for a computational platform designed to run a 
software set for fifty years (or more), we can assume that the platform is 
“static”, or relatively unchanging.  This is to say  that the machine could 
have a basic von Neumann structure that will support a general operating 
system kernel. While over time some speed or miniaturization may take place, 
the general specification remains backward compatibility.  Therefore, my 
2058 Methusatech computer will be designed to run 2008 code and vice versa, 
which is done by adhering to platform standards. Elements which are 
comprised of moving parts, or parts that incur wear are built to be easily 
replaced.  Although this is designed to drastically scale back large-scale 
production of electronic components. new units will ultimately be made, and 
a market for replacement parts will also be needed.  

What Operating System?
The choice for an OS standard could be a fictional 
new system designed to support the platform, but the various flavors of “-
ix” (Unix, Xenix, Linux) have been around for at least thirty years, and 
seems to be the logical candidate for the platform.  The degree of 
stabiility of availabilty, scalability, open source standards, 
interoperability between platforms, and user base make it a good choice for 
the progenitor of the MehusalOS operating system.  There are also older 
systems, such as CP/M, OS9 (not the Apple version), but the choice of an -ix 
system reflects a cultural continuity of nearly forty years.

Software-centricity
So far, we have been talking about the creation of a “static” platform and 
scalability of operating system, but what does that do to the 
production/development culture of the platform?  It creates one of 
“software-centricity”, that is, a decentralization  of hardware manufacture 
beyond that of basic information appliance, and places all of the emphasis 
upon that of software development.  First, stasis of hardware standards 
means that programmers are able to delve more deeply into the intricacies of 
the platform without concern for standards change. The general internal 
structure, libraries, etc. will remain   stable or easily upgradeable on 
older platforms. 

A great example of innovation through limitation is shown through the “demo 
scene”, or a community of artists, musicians, and graphics hackers, who make 
small graphics/audiovisual demos, with the largest community in Europe.  
Although a full discussion of the “scene” is beyond our scope, and their 
community operates on currently evolving platforms as well as “dead” ones, 
the 2K, 4K demo challenges exemplify a model for software-centric design.  
In this model, the coding team is challenged to create a fully functional 
demo that exists below the 2048 (2k) or 4096 (4k) byte size constraint.  As 
hard drives what reach into the terabyte range become more commonplace (and 
this may seem quaint by 2020), acceptance of limitations of software size 
challenges the programmer to extract all possible work from every cycle. It 
eliminates problems of “bloatware” as evidenced by Windows Vista, but the 
user may still accumulate flash drives of programs and data, which is not 
surprising.

A Movement towards Elegance
As mentioned before, the creation of a static platform for general computing 
eliminates the frustrations of continual upgrading and shifts the primary 
focus in computer production to that of software development.  This creates 
a culture of elegance in coding, and craftsmanship amongst coders unseen 
since the days of personal computing of the 1980's.   One could even say 
that a meritocratic market could arise for “code magicians” who are masters 
of programming within constraints.  This is evident in programs written in 
the latter days of platforms like the Atari 2600 game platform like BASIC 
Programming, Visicalc for the Atari 800 computer, and even many applications 
for contemporary mobile phones.  Once again, the matter is that we are 
talking about a platform for essential computation, and not a complete 
replacement for the contemporary computer, although hopes might be that it 
may become one.

Kicking Moore out of the Temple
In 1965, Gordon Moore wrote his seminal article that would create the 
foundation of “Moore's Law” or that the density of transistors in integrated 
circuits doubles every two years.  While this has led to a similar increase 
of the power of computation, it has also created a technoindustrial 
treadmill where there is a politics of fear based on the postwar design 
agenda of planned obsolescence which has translated into the culture of 
personal computation.  From so many personal experiences as an educator, and 
noting the annual software upgrade cycle, the fear of not using the latest, 
greatest hardware connotes one's relegation into the “dustbin of history”, 
as the utility/viability of computation is often linked to the desire for 
more (computing) power and the fear of being linked to technologies 
victimized by planned obsolescence.  While taking our current polemic 
desires to make making Moore's law irrelevant, it also does not ignore the 
fact that what what is being suggested is a genre as well as a paradigm.  
Pragmatically speaking for the foreseeable future, there will be desktops 
and larger laptops, but this paper suggests standardization of even lighter, 
inexpensive computing devices as is evidenced by the sub-notebooks of the 
late 2000's.   

Use What You Need, Hardware or Software
Returning to the analogy of the old “basic four” programs, we are 
questioning the necessity of large media production machines and gaming 
desktops. To reiterate, the 50-Year Computer is not necessarily a 
replacement to the large desktop, but a small, set platform for personal 
computing.  Under this paradigm, one could say that one is using “what they 
need” as opposed to “what they want”.  Those not interested in the mindset 
of necessity will “want” of the larger machines.  

This is not to say that this new computational model is any less flexible 
than an iPhone.  While the hardware remains static (or relatively so), the 
software-centric model discussed before creates more tightly coded, 
“concise” programs. This stresses specific utilities, as the computer is not 
limited by any means to the “Basic Four”.  In addition, there will 
undoubtedly be some operating system updates, as well as firmware, given the 
two maxims of hardware consistency and forward/backward compatibility.  
Conciseness also means that each program will be more specific in their 
function, akin to contemporary small “apps” for portable devices.  The 
interoperability of these apps is also part of the way around the 
limitations of the platform, and this is related to programmatic 
distribution of tasks.

Programmatic Distribution of Tasks
So, how does one deal with limitations of a static platform in regards to 
tasks that grow beyond the capacities of the machine, such as expanded 
spellchecks, etc.  This is done through a set of data standards and 
programmatic distribution of tasks.  One example is that of invoking a 
spellchecker if needed, even matters like image filters, as separately 
invokable routines.   This is only a partial solution to the limitations of 
the system, as chaining together multiple data streams across vast 
archipelagoes of sub-applications will become unnecessarily cumbersome.  
Again, the 50 Year Computer is designed to address essential computing, is 
positioned against Moore's Law, and the aforementioned task-sharing is 
envisioned as a potential method for squeezing every last cycle's potential 
from the machine.

Scalability
All things being said, this writer is not a complete romantic, and 
understands that over the fifty years of production of the hardware between 
2008 and 2058, there might be some upgrading of the platform. In such a 
case, the essential aspect of the platform is that of scalability and 
backward-compatibility. This would be done through retaining the core kernel 
of the operating system and consistent general hardware architecture.  As 
envisioned, the only sacrifice that the 2008 Methusatech user would have as 
opposed to the user of the new 2058 model is less internal RAM (they could 
use external non-volatile RAM), and perhaps some speed.  The hope for the 
system is that the operating system and hardware would be upward and 
downward scalable so that a relatively consistent user experience would 
exist between the 2008 and 2058 models of the machine.

Not Cloud or Network Computing
During the first fifteen years of the World Wide Web, on at least two 
occasions there have been alternate models to the desktop in the form of the 
Network Computer and Cloud Computing.  The former, started in the 1960's and 
attempted as a paradigm in the 1990's, was a method in which programs would 
be served remotely, allowing for the majority of storage to be done through 
servers.  The NC, was in effect, a “really smart terminal”.  Conversely, 
Cloud Computing borrows from the distributed models of SETI Online, which 
allows millions of users to analyze radio telescope data to determine 
correlations for possible intelligent life.  Machines are clustered to 
distribute their computational power through the creation of “clouds” of 
processors.  Although the 50-Year Computer is surely not an “NC”, and it 
might be technically able to be used as a cloud machine, its function is 
envisioned as a single-person information appliance for the execution of 
essential personal computing. 

50-Year Computer: Not a Blow to Innovation
Recently, while discussing this concept, I was challenged that this idea was 
against innovation, and this is patently not the case.  It merely suggests a 
different paradigm to cultures of innovation.  For example, a timeline was 
laid out for me outlining increases in resolution, the addition of color, 
increases of power mitigating increases in amenities, to which I asked to 
what ends these advances served.  “It answered the question of what people 
want...”, he said.  Once again, we are faced between desire and utility in 
the face of a Fordist production scheme that increasingly puts more toxic 
landfill in the 3rd World and creates an endless cycle of fear, desire, and 
obsolescence, and the 50 Year Computer refutes this.  On the other hand, it 
creates a new culture of innovation in the software sector, which is a 
paradigmatic shift from contemporary models.  These models of software and 
hardware production are patently unsustainable, and by allowing us to think 
of operating “antique” computers, we also antiquate the notion of planned 
obsolescence, even though proposing a radical new paradigm may seem ironic 
in this context.

Conclusion
The 50-Year Computer, or something like it, is not a “modest proposal” that 
merely dismisses the computational culture of Moore's Law, and 
technodeterminism, although it does challenge it.  From my experience in 
computing for the past thirty years, much of what I call “essential” 
computing still revolves around functions served by basic office software, 
with the only addition being functionality for the Web.  Although 
contemporary computation allows for wonderfully rich and engaged 
experiences, it's arguable as to whether many more necessary programs, 
including aforementioned Web technologies have been created in the past 
thirty years.  From this, the 50-Year Computer is a  statement is designed 
for reflection upon the grossly inefficient, wasteful, and unsustainable 
computational culture extant in the technological world.  In proposing the 
50 Year Computer, I suggest many things; a “Model T” computer that performs 
utilitarian functions for decades, reduction of energy consumption, and 
waste streams, as well as maintaining occupation of labor by redistributing 
it to software production.  This is not to say that the 50 Year Computer 
will eliminate the desktop, as it is unreasonable to expect to eliminate an 
entire computational culture of scale merely by wishing it.  The  
proposition of this missive is to suggest alternatives for computation that 
could offer greater sustainability, affordable utilitarian computing to the 
global masses, and cultures of software artisanship.  What is clear is that 
as the 2010's approach, if we are to continue using computational devices, 
we must consider alternate methodologies that take in account the use value 
of computation to world masses as well as general sustainability.