Requirements Analysis

Requirements Analysis is one of the Systems Engineering tools to do the design and optimization of a system.

A system as a whole has a statement of what we want it to do. This is in the form of goals, values, mission description, performance requirements and such. It also needs ways to measure how good a given design is. These can be things like "minimum cost", "minimum waste output", and "maximum efficiency". These statements and measures are often in verbal form. Requirements Analysis is the process of putting them in numerical form, breaking them down to more detail, and assigning them to functions who will perform them. The assignment ensures that somewhere in the system the top level goals are met. At the most detailed level a subset of the requirements are assigned to a single function box. This now becomes the conditions that a specific engineering design needs to meet.

Some stability is needed in requirements so that the design work can proceed. How to document the results of this analysis task and when to change it needs to be determined.

Global Village System Identification
The first step is to identify what is the system we are working with, and what is outside the system (the environment). The Global Village Construction Set is a set of machines and technologies, but are not sufficient in themselves. You also need land on which to use the machines, people to operate them, and information so the people know what to do. These combine to form a Global Village System (GVS), which as a whole serves the purposes for which it is designed.

The environment in which the system operates is everything outside the boundary of the system, in other words the entire rest of the universe minus the system. Among the outside entities the system can interact with are more copies of itself. You can define inputs and outputs that flow across the system boundary. As a time series, at first more inputs are required, otherwise the system would be empty. During it's operating life both inputs and outputs will occur. If the system reaches an end of life and is dismantled, then outputs will outweigh inputs until nothing is left, and all parts are disposed of properly.

Multiple versions of the system can exist during design as alternatives. Once design is completed, multiple versions can exist to account for differences in number of people, land area, climate, geology, starting funds, diet preference, and surrounding region economic state and technical infrastructure. As a starting point, a single nominal "design point" can be established, from which versions can be derived.

Source Statements and Measures
This is a list of source statements and measures for what the Global Village System is supposed to do, and then a restatement in a measurable form.

From OSE Mission:


 * "remove material scarcity" > Provide a surplus of free time and useful outputs from the system after accounting for internal needs and inputs from the environment.


 * "harmonious coexistence between natural and human ecosystems" > The system maintains and improves itself with a minimum of non-renewable inputs and waste outputs.


 * "land stewardship, resilience, and improvement of the human condition" > This is met by the above two statements.


 * "everybody's needs are met" > Provide sufficient surplus against times when a person cannot contribute due to age, health, or other reasons and to account for mechanical failure or natural variations.


 * "replicable...communities" > Maximize the ability of the system to copy itself, as opposed to building from scratch as the first copy has to be, and allow communities to grow in number by seeding or fission (in the biological sense).

From OSE Specifications:

Not every specification is listed here when they lead to the same requirement.


 * Specifications 2: "local food systems...local digital fabrication" > Minimize transportation energy and resources.


 * Specifications 3: "5-10x cheaper than buying" > Minimize hardware cost including fabrication labor.


 * Specifications 4: "Components of the GVCS function as interchangeable modules" > Design for optimum modularity and commonality


 * Specifications 10: "Our products are designed for a lifetime of use." > Optimize life cycle inputs and outputs


 * Specifications 11: "Our products substitute common resources for less common resources." > Minimize scarce and narrowly distributed resources


 * Specifications 12: "Our digital fabrication equipment can flexibly produce a huge variety of products" > Maximize downstream process capability.

Top Level Requirements

 * Provide a surplus of at least 1.2 times labor input, in other words (total output/total input) > 2.2 times input.
 * Rationale: The ratio of total people to employed persons in the United States is 2.2, and thus each employed person supports 1.2 additional people. The GVS should be at least as productive as that.  Employed persons and GVS labor input both do not account for non-job activities such as household chores.


 * Operations and maintenance of the system less than 100% of system capacity.
 * Rationale: In order to maintain and improve itself, the GVS as a whole needs to spend less than all it's outputs on itself. For example, if you spent 100% of your labor time growing food, you have no time left to do anything else.  Similarly other components need to used less than 100% to have some capacity left over for improvement and growth.


 * Minimize non-renewable inputs to less than 100% relative to conventional equivalents.
 * Rationale: The Earth has a finite supply of non-renewable resources. Thus for long-term sustainability use of those resources should be minimized.  The GVS should use as little as possible, but at least below the average of the surrounding community who are not using the GVS.  If the GVS can produce an equivalent surplus output, for example biodiesel for sale, then use of petroleum-based inputs no longer falls into the "non-renewable" category, and use can then be negative (outputs exceed inputs).


 * Minimize waste outputs to less than 100% relative to conventional equivalents.
 * Rationale: Similar to the previous rationale, the Earth has a finite ability to absorb waste outputs which degrade the environment. Therefore the GVS should produce as little as possible, but at least below that of the surrounding community.  If you can take in a local waste product, for example paper that would otherwise be burned or go to a landfill, and use it to make something useful, like papercrete blocks, that would count as negative waste, since you are reducing total waste in the environment. Design of flows and functions should use wastes from one step as input to another.  For example human and kitchen wastes become fertilizer inputs.


 * Maximize the percentage of system components that can be efficiently produced by the system.
 * Rationale: A certain amount of outside inputs are required to set up the first copy of a GVS in a given area. By producing it's own parts, additional copies will require less inputs.  Copies can be made from seeding or fission.  Seeding is by a GVS providing a minimal "starter set" to set up another GVS, which then grows to a full version.  Fission (or mitosis biologically) is by a GVS growing to twice it's original size, then splitting into two complete copies.  The "efficiently produced" means relative to getting it from an outside source.  For example you may need a computer to control automated fabricators.  CPU and memory chips, and hard drives are produced in highly specialized, expensive factories.  So they cannot (at present) be efficiently produced locally.  Structural metals, however, can be fabricated with relatively simple equipment.  The useful measure then may be the ratio of (amount of that component produced/equipment needed to produce it).


 * Minimize total transportation energy and waste products.
 * Rationale: Shipping items long distances adds cost, consumes resources, and adds waste products to the environment. It can also degrades item quality in some cases, like for food.  So other things being equal, it is preferable to grow food and manufacture items locally.  But if local production is less efficient, that may outweigh the transportation savings.  For example, keeping a chicken coop warm through the winter might use more energy than shipping them from a warm climate.  So total energy use has to be considered, not just transportation energy.


 * Minimize hardware cost including labor = significantly less than purchased equivalent.
 * Rationale: Make/Buy analysis is a normal part of engineering. Some parts are simply too hard or expensive to make compared to buying. When counting labor cost, a finished assembly or total machine may end up more expensive than a purchased unit.  The design goal for a GVCS machine should be significantly less cost to allow for other factors, like maintenance time, being somewhat worse.


 * Optimize modularity and commonality.
 * Rationale: The source statement references the GVCS, which are the machines. Here it is widened to the entire GVS.  For example, building components made modular and common sizes so they will fit together.  This is already done in existing USA construction, where materials are often in multiples of 2, 4, and 8 feet.


 * - Modules reduce downtime for maintenance - swap a module and get back to operations.
 * - Shared modules also allow less total hardware when machines are not used 100% of the time. One power module can be shared among several machines, for example.
 * - Common parts also reduces inventory and training.
 * - The requirement is to "optimize", not "maximize". A 50kW power unit for a tractor is too big for a 200W drip irrigation pump.  So several sizes are acceptable, as long as they are standardized.  Optimum is determined when it gives the best result for the total system.
 * - Adapting to changing needs can be handled by adding more copies of common modules, and trading off to other Villages when no longer needed.
 * - Standard connections between modules allow swapping different designs. For example solar and wind vary depending on location, but if they use the same connections for output power and automated control they can be swapped as needed to suit the location.


 * Optimize long term life cycle inputs and outputs.
 * Rationale: The GVS is intended to support human communities for the long term. Thus the initial cost and inputs has to be weighed against continuing costs, maintenance, inputs and outputs, with more weight towards the latter.  The appropriate "long term" time scale should consider civilizational changes such as population and technology progress.  For example, designing a solar panel to last a thousand years is excessive when the technology is improving 3% a year.


 * Minimize scarce and narrowly distributed non-renewable inputs.
 * Rationals: Scarce is measured by whether there is enough of the resource for the entire world to use it at GVS levels for the long term. The GVS should not over-use such resources such that it would run out.  Narrowly distributed is measured by percent of world population within acceptable distance of a source.  Sunlight at some level is everywhere, so would score 100%. A specific mineral only available from one mine would score very low.


 * Maximize the weighted fraction of later generation process tools and products.
 * Rationale: There are a finite number of manufacturing processes. If the initial GVCS tools can be used to make second and later generation tools that implement more processes, then a wider range of final products can be made.  Weighted fraction means what percent of the products by some measure (weight, value, number of people who use it) can be produced.  Absolute number of processes you can implement is not as important since many of them can be substituted, and some are rarely used.

Measures of Effectiveness
With multiple requirements listed above, there must be some way to decide among different designs. Each design will vary in how well it meets different requirements. The way to compare them is to have a scoring system. This starts with a measurable value of the design, for example tons/year of waste C02, which would be a part the "minimize waste outputs" requirement above. You make a scoring formula. For example 0 tons CO2 = score of 100, 1000 tons CO2 = score of 0, with varying values in between. Then you make a total score formula for how to add up the the component scores of a given design. This allows more important measures to be given heavier weight in the total score.

Then choosing between different designs becomes a matter of scoring each one and seeing which is highest. In the early stages of design you will have some uncertainty in the measures, and so uncertainty in the final scores. If a given design is clearly worse in final score when you account for the uncertainty, you can stop working on it and concentrate on the high scoring ones. If the uncertainties get small and two or more score close together, at some point it is not worth the effort to reduce the uncertainties further. Just pick one and finish it's design.

Measures should be consistent across the entire system. Otherwise you end up optimizing individual parts at the expense of the rest of the system. Choosing what measures to use and how to score them is subjective. It is based on what people think are good, bad, important, or not important. Once chosen, however, the scoring system should be numerical and objective. Everyone can see how the result was reached, and it should exclude decisions made because one person liked a given design more regardless of it's actual merits.