Skip to content

Syntax and semantics

In this section, we introduce in an informal way the syntax and semantics of the various ESL concepts. The syntax and semantics is presented for a selection of ESL concepts from the full ESL Parsing Expression Grammar (PEG) definition.

Tip

The most up-to-date grammar definition in PEG format can always be found over at the language's tools GitLab project in the grammar.pest file.

Components

Many engineers experience difficulties in defining and mapping system, function and variable decompositions to each other1. Therefore, ESL only allows for the definition of a single decomposition: a system decomposition tree (product breakdown structure) consisting of components. The components represent parts of the system.

Each component must have a definition and must be instantiated as a child component within a (higher-level) parent component. ESL has the built-in world component which is the root of the decomposition tree. The world is not part of the system but represents the environment in which the system operates, which allows user to describe interactions between the system and its environment. The component tree (system decomposition) forms the central structure of an ESL specification. All ESL statements are placed within the component tree. For example, shows a world containing a single component drive-mechanism which is an ElectricalDriveMechanism, which in turn is empty, i.e., it does not contain any child components (yet).

1
2
3
4
5
6
world
  component
    drive-mechanism is an ElectricalDriveMechanism

define component ElectricalDriveMechanism
  empty

In the listing below, two child components are instantiated within the definition of the ElectricalDriveMechanism. As such, drive-mechanism is no longer empty but consists of electric-motor, which is a BrushlessMotor, and power-source, which is a Battery. In turn, one could define and instantiate child components within the BrushlessMotor definition and the Battery definition to add additional layers to the system decomposition.

world
  component
    drive-mechanism is an ElectricalDriveMechanism

define component ElectricalDriveMechanism
  components
    electric-motor is a BrushlessMotor
    power-source is a Battery

define component BrushlessMotor
  empty

define component Battery
  empty

Since the system decomposition tree is the central structure in an ESL specification, one takes a system centered modelling perspective in writing an ESL specification. That is, components are viewed as black boxes with inputs and outputs erden_review_2008. This may seem to be in contradiction to scholars who advocate 'solution-free' modelling, such as 2 and3. In solution-free modelling, one usually takes a process centered modelling perspective, in which processes are viewed as black-boxes with inputs and outputs erden_review_2008. However, in an ESL specification, a component may represent a conceptual thing that must do something within a process. A conceptual component does not impose a solution. For example, in the example above, one could instantiate a general-drive-mechanism instead of an ElectricalDriveMechanism in case one does not want to specify the working-principle of drive-mechanism.

Types and variables

In engineering design, it is common practice to describe a design in terms of variables. ESL enables the user to declare variables within components. The variables may represent flows from the interaction basis 4, such as electrical energy and information, or properties (attributes) of components, such as length, weight, cost, and reliability.

Each variable must have a type, which needs to be defined by the user. The type of a variable provides additional information about the nature of that variable. For example, whether the variable represents a solid or a liquid material flow. Moreover, the variable types enable consistency checks as variables are passed along multiple levels of the system decomposition tree. ESL supports the mechanism of actual and formal parameters to pass variables across levels of the decomposition tree, which is common in programming languages.

In the following ESL example, the type electrical-energy-flow is defined in line 2. In line 6, the variable power is declared which has type electrical-energy-flow. This actual variable is passed along as an argument to components electric-motor and power-source (lines 10, 13). Upon instantiation of BrushlessMotor and Battery the formal parameters input-power (line 17) and output-power (line 21) are replaced by the actual variable power.

define type
  electrical-energy-flow is a real

define component ElectricalDriveMechanism
  variable
    power is an electrical-energy-flow

  components
    electric-motor is a BrushlessMotor with arguments
      * power

    power-source is a Battery with arguments
      * power

define component BrushlessMotor
  parameter
    input-power is an electrical-energy-flow

define component Battery
  parameter
    output-power is an electrical-energy-flow

Domains

Semantically a type denotes the kind of values that a variable of that type can assume. For example, a real number. A domain denotes the set of possible values that a variable of a certain type may take. For example, all non-negative real values. In math, the domain of continuous data is usually indicated by lower and upper boundaries. For discontinuous data enumerations are often used that list the set of possible values.

In ESL, the domain of continuous data may be defined by a closed interval, an open interval or multiple disjunct intervals. Closed intervals can be defined in the following manner:

define type
  foo is a real of at least 0.0 and at most 10.0

That is, the definition of a type is extended with a lower and upper bound. In this case, a variable of type foo can only assume real values of at least 0.0 and at most 10.0. By omitting either the lower or upper bound one can create an open interval. For example:

define type
  faa is a real of at least 0.0

indicates that variables of type faa can assume real values of at least 0.0. One can define two disjunct open intervals by using an or construct. For example:

define type
  fuu is a real of at most 0.0 or at least 10.0

indicates that a variable of type fuu can assume real values of at most 0.0 or least 10.0. In general, the or construct can be used to define multiple (open) intervals. For example:

define type
  fee is a real of at least -1.0 and at most 0.0 or at least 10.0 and at most 20.0 or at least 100.0

indicates that a variable of type fee can assume real values of at least -1.0 and at most 0.0 or at least 10.0 and at most 20.0 or at least 100.0. In other words, the domain of a variable of type fee is defined by two closed and one open interval that are disjunct. The ESL compiler checks for empty intervals, overlapping intervals and inconsistencies in sub-intervals.

An interval spanned by domain boundaries is different from an interval spanned by a set of design-requirements or constraints. An interval spanned by domain boundaries indicates all possible values that exist while an interval spanned by a set of design-requirements or constraints denotes all values that are desired or feasible. Therefore, an interval spanned by a set of design-requirements or constraints with respect to a certain variable must overlap with the interval spanned by domain boundaries of the type of that variable. Otherwise, the design space is infeasible.

The domain of discontinuous data is usually defined by at set of distinct values. In ESL, enumerations can be used to define the set of distinct values that a discontinuous variable may take. For example:

define type
  fii is an enumeration of red, blue, 1, 2

indicates that a variable of type fii may assume a value of red, blue, 1 or 2. The values of an enumeration may have heterogenous types.

Units

By default the implicit [-] unit is assumed for all values in domain boundaries, enumerations and constants. However, as explained in LEP-0003, in feasibility studies units are required to verify whether a comparison between a variable and a literal value, or between two variables makes any sense. Such a verification remains impossible of no units are attached to the values within the domain of a variable. That is, an upper bound of 10 m/s is different from an upper bound of 10 km/s. Therefore, units can be specified after the values used within domain boundaries, enumerations and constants.

For example, the following listing shows several good and bad examples of how to define units within domain boundaries:

1
2
3
4
5
6
define type
  speed is a real with units m/s, km/h
  human-speed is a speed of at least 0.0 [m/s] and at most 10 [m/s]  # Ok
  car-speed is a speed of at least 0.0 [km/h] and at most 220 [km/h] # Ok
  boat-speed is a speed of at least 0.0 [knots] and at most 20.0 [knots] # ERROR,  unit not allowed.
  air-speed is a real of at least 0.0 [km/h] and at most 960.1 [km/h] # ERROR, unit not allowed.

That is, first one has to define a type that represents the physical quantity of interest and its allowed units. Here being speed with units m/s and km/h. Subsequently, one can define types that represent subdomains of type speed. For instance, types human-speed, car-speed and boat-speed denote subdomains of type speed. Here boat-speed will yield an error as the used unit knots is not allowed for the physical quantity speed. Type air-speed will yield an error as well as the build in type real has the default [-] unit.

Similarly, one can specify units after values used within constants. For example, in the following listing the physical quantity acceleration is defined with unit m/s^2. The type gravitational-constant represents the value of 9.81 m/s^2 within the acceleration domain.

1
2
3
define type
  acceleration is a real with unit m/s^2
  gravitational-constant is an acceleration equal to 9.81 [m/s^2]

Enumerations are slightly different as the values within an enumeration may be heterogenous in type and therefore do not span a sub-domain of a parent type. As such, the unit of a value may be specified directly after the value. For example:

define type
   engine-outputs is an enumeration of 10 [Nm], 20 [Nm], 30 [Nm]

Properties

In engineering design it is common practice to define properties of components, such as the total weight, maximum length, and cost. Therefore, ESL contains the property keyword. For example, in the code block below, the property keyword is added behind the length parameter declaration (line 14). By doing so, the variable drive-length (line 2), which is mapped onto formal parameter length as it is an argument of drive-mechanism (line 10), is explicitly marked as being a property of component drive-mechanism. A variable can be a property of at most one component.

define type
  geometry is a real with unit [mm]

world
  variable
    drive-length is a geometry

  component
    drive-mechanism is an electric-drive-mechanism with argument
      drive-length

define component electric-drive-mechanism
  parameter
    length is a geometry property

Bundles and groups

Many programming and modelling languages contain container data types that enable users to pass along information in packages. Similarly, ESL contains bundles and groups. Bundles are a special variable type that enables users to declare multiple variables at once. For example, in the listing below the type electrical-power-bundle (line 5) is a bundle of two fields. One named U which is a voltage and one named I which is a current. The declaration of variable P of type electrical-energy-bundle at line 11 implies te declaration of the variables P.V and P.I.

define type
  voltage is a real
  current is a real

  electrical-power-bundle is a bundle of
    * U is a voltage
    * I is a current

world
  variable
    P is an electrical-power-bundle

Groups can be used to (re)group already declared variables. A variable may be part of multiple groups. For example, in the listing below the variables P1 and P2 of type electrical-power-bundle are declared at line 11. This implicitly implies the declaration of variables P1.V, P1.I, P2.V, and P2.I. At lines 14 tot 20 these variables are re-grouped to form a group of voltages and a group of currents.

define type
  voltage is a real
  current is a real

  electrical-power-bundle is a bundle of
    * U is a voltage
    * I is a current

world
  variables
    P1, P2 is an electrical-power-bundle

  variable-groups
    voltages is a group of
      * P1.V
      * P2.V

    currents is a group of
      * P1.I
      * P2.I

Verbs and prepositions

The many synonyms of verbs in natural language may cause ambiguity in system specifications deng_function_2002. Therefore, ESL enforces users to define all verb-preposition combinations that are allowed in ESL function specifications. It is advised to use verbs from the functional basis of 5, which all have a distinct definition.

Needs

Needs are informal statements on what is desired6. A need specification within ESL must start with a reference to an instantiated component or a reference to a declared variable. The component or the variable is the subject of the need. All words that follow the subject are unconstrained, i.e., a user may write pure natural language. A need may, for instance, describe desired functionality.

Function specifications

The goal- and transformation-function grammars by 7 fit well within the system centered modelling perspective of ESL as goal- and transformation-functions are formulated in terms of components and variables. The concepts of goal- and transformation-functions, as presented by 7, are therefore woven into ESL.

Goal-function specifications denote the purpose of components with respect to other components within the system. ESL enables the specification of goal-functions in requirement form and constraint form. Goal-requirements state what a to be realized component must do, while goal-constraints state what an existing component (design) does. The listing below shows an example goal-function in requirement form (gr-0) and in constraint form (gc-0). Note that the only difference between these two forms are the verbs must and does.

1
2
3
4
5
goal-requirement
  gr-0: Battery must provide power to lamp

goal-constraint
  gc-0: Battery does provide power to lamp

The general form of a goal-specification literal (sentence) is defined in the listing below, which presents the grammar in Parsing Expression Grammar (PEG) notation:

goal_line = {
    some_component ~              // Active/sending component
    kw_aux ~                      // "does" (constraint) or "shall"/similar (requirement)
    some_verb ~                   // I.e. "provide"
    some_preposition ~            // I.e. "to"
    some_variable ~               // Variable indicating a flow from active to passive
    (some_variable ~ kw_enum)* ~  // Additional variables
    some_component ~              // Passive/receiving component
    with_subclauses?              // Optional additional clauses
}

The first some_component denotes the active component, i.e., the component that actively fulfills what is described by the goal-specification. The kw_aux must be one of shall, must, should, could, or won't. The VERB and its associated PREPOSITION denote the action performed by the active component. The argument-references is a list of arguments, i.e., either variables or parameters, that quantify the goal specification. The second COMPONENT-NAME denotes the passive component, i.e., the component that passively contributes to fulfilling what is described by the goal-specification. The subclauses are optional and state additional design specifications that are subordinate to the main clause. For example, goal-requirement g1 in the listing below states that active component drive-mechanism-d must transfer torque to passive component pump-p with a steady-state-torque equal to 40 Nm (indicating a single operating point) and reliability of at least 95%. T

1
2
3
4
goal-requirement
  g1: drive-mechanism-d must transfer torque to pump-p with subclauses
      * c1: steady-state-torque must be equal to 40 Nm
      * c2: reliability must be at least 95 %

The grammar of subclauses is defined in the listing below:

1
2
3
4
5
subclauses ::=
  "with" "subclauses" \n
  { "*" SUBCLAUSE-NAME ":"
    comparison-rule-line \n
  }+

The subclause section starts with the keywords with subclauses followed by a new line and one or more subclauses, each consisting of a SUBCLAUSE-NAME and a comparison-rule-line. The comparison-rule-line is defined in section Design Specifications.

Transformation specifications describe the internal conversion processes within a component. ESL provides for the specification of transformation-requirements and transformation-constraints. Similar to the goal specifications, transformation-requirements describe what a desired component must do internally, while transformation-constraints describe what an existing component (design) does internally. The listing below shows an example transformation-function in requirement form (tr-0) and in constraint form (tc-0). Note that again the only difference between the two forms are the verbs must and does.

1
2
3
4
5
transformation-requirement
  tr-0: must convert power to light

transformation-constraint
  tc-0: does convert power to light

The grammar of a transformation literal (sentence) is defined in the listing below.

1
2
3
4
5
6
transformation-literal ::=
  ( "does"           % constraint
  | auxiliary-verb ) % requirement
    VERB argument-references
    PREPOSITION argument-references
  ( \n | subclauses )

Transformation literals are to be specified in the body of the component definition. Therefore, a transformation-literal starts with the verb does, in case one wishes to specify a constraint, or starts with an auxiliary-verb, in case one wishes to specify a requirement. The transformation literal subsequently consists of a user-defined VERB with argument-references representing the transformation input parameters, a user defined PREPOSITION followed by argument-references representing the transformation outputs, and optionally subclauses.

For example, transformation-requirement in the listing below states that each instantiated BrushlessMotor must internally convert power into torque.

1
2
3
4
5
6
7
define component BrushlessMotor
  parameter
     power is an electrical-energy-flow
     torque is a mechanical-energy-flow

  transformation-requirement
    t1: must convert power into torque

Behavior specifications

Modern engineering systems, however, are usually dynamic in nature. That is, they change behavior and/or state in response to stimuli such as operator input or changes in their environment. As such, there is a strong need to be able to specify when a system should exhibit what behavior. In other words, there is a strong need to be able to specify dynamic bounds on flows.

In the literature, system behavior is often modelled in conjunction with system functions and system states. See, for example, the Function-Behavior-State model 8, the Structure-Behavior-Function model 910 and the requirements engineering book of Hull11. These models all have their own definitions of behavior, function and state.

In light of the literature and the flow based foundation of ESL, function, behavior and state are defined as:

  • Function: everything what a 'thing' must be able to do, i.e., all transfers and transformations of flow which a 'thing' must be able to perform.
  • Behavior: everything what a 'thing' must do in response to stimuli, i.e., when which flows need to be transferred and transformed in what quantity in response to stimuli. Stimuli are changes in (un)controllable flows (state changes).
  • State: description of the behavior of a 'thing' given certain bounded stimuli, i.e., what flows are being transferred and transformed in what quantity given certain bounded stimuli. A state change happens when a stimulus crosses the defined boundaries. For example, a sensor signal (information flow) switches from high to low.

Note

When following the state definition above, a state is a representation of the values of a set of flows within certain bounds. In ESL all flows are represented by variables. Bounds are variables as well. A state in ESL is specified as a variable bounded by other variables.

In the literature one can find many boilerplates for behavior requirements, see for example 9, mavin2009easy and12. In general, boilerplates have a when \< A > then \< B > like structure. The when keyword indicates a set of conditional statements A that indicate under which conditions requirement set B must be satisfied.

Furthermore, behavior requirements usually describe a set of alternative cases. Such requirements follow a when \< A1 > then \< B1 > or ... or when \< An > then \< Bn > structure. The order of the cases in such a requirement is usually meaningless. That is, the order of case specification does not impose any precedence constraints.

Following the literature, the behavior syntax as shown in the example below is proposed. In this example, a behavior requirement named door-state is specified. The requirement comprises the two cases door-is-closed and door-is-open. The case order has no semantics. Each case is composed of when-clauses and then-clauses. The when-clauses state the conditions under which the then-clauses apply. That is, when condition c1 holds then requirements r1 an r2 must be satisfied. When condition c2 holds requirements r3 and r4 must be satisfied.

behavior-requirement
  door-state:
    case door-is-closed:
      when
        c1: door-sensor-voltage is greater than 0.9 V
      then
        r1: door-open must be equal to False
        r2: door-closed must be equal to True

    case door-is-open:
      when
        c2: door-sensor-voltage is at most 0.9 V
      then
        r3: door-open must be equal to True
        r4: door-closed must be equal to False

In writing behavior requirements one should strive for complete and deterministic requirements. Complete implies that the conditions cover the complete range of stimuli. Deterministic implies that only one case may apply simultaneously. Note that the example above is complete and deterministic as conditions c1 and c2 cover the full range of possible values of door-sensor-voltage and can never hold simultaneously.

Fall-back cases

In striving for completeness, exhaustively specifying all possible cases may become a cumbersome and time consuming task. As such, it is desirable to be able to specify a fall-back case. A fall-back case is similar to the else construct used in many programming languages.

A special when no other case applies clause is used to enable users to specify fall-back cases. The example below specifies behavior requirement output-x which contains cases A, B and C. Case C is the fall-back case that applies when neither A nor B applies. In other words, C applies when y is at least 5 and at most 10.

behavior-requirement
  output-x:
    case A:
      when
        c1: y is smaller than 5
      then
        r1: x must be at most 5
    case B:
      when
        c2: y is larger than 10
      then
        r2: x must be at most 10
    case C:
      when no other case applies
      then
        r3: x must be equal to 0

Note that the when no other case applies when-clause is a short-hand notation for negation of the when-clauses of case A and B.

Behavior, goal and transformation specifications

Behavior specifications are complimentary to goal and transformation specifications. In fact, the addition of behavior requirements allows for more natural transformation specifications. For example, assume we have the following three goal-requirements:

1
2
3
4
goal-requirements
  send-control-signal: motor-controller must send motor-control-signal to motor
  provide-power: power-supply must provide power to motor
  provide-torque: motor most provide torque to pump

which are dependent on one another. In ESL version 1.0 one has to define a transformation-requirement convert-power, as shown below, to ensure that goal provide-torque has a functional dependency with goals send-control-signal and provide-power.

1
2
3
4
5
6
7
8
define component electric-motor
  parameters
    motor-control-signal is an information-flow
    power is an electrical-energy-flow
    torque is a mechanical-energy-flow

  transformation-requirement
    convert-power: must convert motor-control-signal and power into torque

Physically, however, motor-control-signal is not really transformed into torque by motor. As such, transformation requirement convert-power may seem flawed by readers.

By adding a behavior-requirement, as shown in the listing below, one can separate the physical function of transforming power into torque from the logical behavior that motor must only provide torque when needed.

define component ElectricMotor
  parameters
    motor-control-signal is an information-flow
    power is an electrical-energy-flow
    torque is a mechanical-energy-flow

  transformation-requirement
    convert-power: must convert power into torque

  behavior-requirement
    motor-behavior:
      case motor-on:
        when
          c1: motor-control-signal is equal to 'on'
        then
          r1: torque must be equal to 60 NM
      case motor-off:
        when
          c2: motor-control-signal is equal to 'off'
        then
          r2: torque must be equal to 0 NM

That is, variable motor-control-signal is no longer an input argument of convert-power. Instead, it is the subject of conditions c1 and c2. This implies that the desired value of torque, which is an output argument of convert-power, is conditionally dependent on the value of motor-control-signal. Consequently, goal provide-torque torque has a functional dependency on goal provide-power and a behavior dependency on send-control-signal. Since power is needed to provide torque and motor-control-signal denotes when torque must be provided.

Syntax

The Behavior EBNF listing below shows the syntax of an ESL behavior section in Extended Backus-Naur Form (EBNF). A behavior section starts with a behavior type keyword, either behavior-requirement or behavior-constraint, followed by a new line. The next line starts with a behavior-NAME, i.e., the identifier of the behavior statement, followed by a colon, a behavior-clause and a new line. Such a pattern is repeated one or more times. A behavior-clause starts with the keyword case followed by a CASE-NAME, a colon and a new line. On the new line one may find a regular-case-clause or a fall-back-case-clause. This pattern may repeat itself one or more times. A behavior-clause may contain at most one fall-back-case-clause.

A regular-case-clause is composed of a when-clause followed by a then-clause. The when-clause states the case conditions and is composed of the keyword when followed by a new line. The new line starts with a CASE-CONDITION-NAME followed by a colon, a comparison-rule-line and a new line. A when-clause must contain at least one condition.

A then-clause states the case design-rules and starts with the keyword then followed by a new line, a CASE-RULE-NAME, a colon, a comparison and a new line. Each then-clause must contain at least one design rule. The then-clauses of the cases within the scope of a behavior-requirement must have the same set of subject variables.

A fall-back-case-clause starts with the keyword sequence when no other case applies and a new line where one finds a then-clause.

behavior-section ::=
  behavior-type \n
  { BEHAVIOR-NAME ":" \n
    behavior-clause
  }+

behavior-type ::=
  "behavior-requirement" | "behavior-constraint"

behavior-clause ::=
  {
    "case" CASE-NAME ":" \n
    ( regular-case-clause
    | fall-back-case-clause
    )
  }+ /* At most one fall-back-case-clause per behavior-clause. */

regular-case-clause ::=
  when-clause
  then-clause

when-clause ::=
  "when" \n
    { "*" CASE-CONDITION-NAME ":" comparison-rule-line \n }+ /* Only 'is' as auxiliary verb. */

then-clause ::=
  "then" \n
    { "*" CASE-RULE-NAME ":" comparison \n }+

fall-back-case-clause ::=
  "when" "no" "other" "case" "applies" \n
  then-clause

Design specifications

Design specifications denote bounds on the values of variables. The listing below shows an example design-specification in requirement form and in constraint form. The verbs must and is indicate the difference between the requirement form (dr-0) and the constraint form (dc-0).

1
2
3
4
5
design-requirement
  dr-0: power must be at least 300 [W]

design-constraint
  dc-0: power is at least 300 [W]

The grammar of a design-rule-line is defined as:

design-section ::=
    design-rule-type \n
  { DESIGN-RULE-NAME ":"
    design-rule-line }+

design-rule-type ::=
    "design-requirement"
  | "design-constraint"

design-rule-line ::=
  comparison-rule-line
  ( \n | subclauses )

comparison-rule-line ::=
  comparison
  { "or" comparison }

comparison ::=
  argument-name
  ( constraint-rule-literal | requirement-rule-literal )

constraint-rule-literal ::=
  "is" compare-op bound

requirement-rule-literal ::=
  auxiliary-verb "be"
  ( compare-op bound | objective )

compare-op ::=
    "smaller" "than"
  | "greater" "than"
  | "not" "equal" "to"
  | "equal" "to"
  | "at" "least"
  | "at" "most"
  | "approximately" /* Only in a requirement-rule-literal */

bound ::=
    argument-name
  | VALUE [ unit ]
  | "t.b.d." [ unit ]

unit ::=
    UNIT-NAME
  | "[" UNIT-NAME "]"

objective ::=
    "maximized"
  | "minimized"

A design-rule-line consists of a comparison-rule-line followed by a new line or subclauses . An comparison-rule-line consists of one or more comparison. A comparison starts with an argument-name followed by a constraint-rule-literal or a constraint-rule-literal. An constraint-rule-literal starts with the word is followed by a compare-op and a bound. An requirement-rule-literal starts with an auxiliary verb and be followed by an compare-op and a bound or an objective.

For example, design-requirement d1 in the listing below states that the value of water-flow-1 must be at least 10 [l/s], while d2 states that water-flow-1 must be smaller than water-flow-2 or water-flow-1 must be smaller than water-flow-3 . The ... denotes line continuation.

1
2
3
4
design-requirement
  d1: water-flow-1 must be at least 10 l/s
  d2: water-flow-1 must be smaller than water-flow-2 or ...
      water-flow 1 must be smaller than water-flow-3

Relations

In engineering design detailed (mathematical) analysis is usually required to ensure that a system will function according to the set requirements. An analysis often consists of defining a mathematical function that requires a set of input arguments and returns a set of output arguments. Based on the values of the input arguments, the values of the output arguments are computed. The values of the output arguments are subsequently (manually) verified against the set requirements.

The amount of information one has one a relation between variables may change during the course of a design project. At the start of a design project, one may only be aware of the existence of a relation. Later, one may obtain additional information that denotes the input-output relations. Therefore, the proposed relation syntax allows for the specification of directed and undirected relations. This enables users to convert undirected relations into directed relations during the course of the design project.

For example, the listing below shows an example specification in which two relations are defined and instantiated following the proposed syntax. At line 2, relation ohms-law is defined which requires parameters current and resistance as input and returns parameter voltage as output. At line 9, relation newtons-second-law is defined which relates parameters force, mass, and acceleration.

At line 21, ohms-law is instantiated as relation r1 which requires arguments I and R and returns argument U. At line 28, newtons-second-law is instantiated as relation r2 which relates arguments F, m, and a.

define relation
  ohms-law
    requiring parameters
      * current is a real
      * resistance is a real
    returning parameter
      * voltage is a real

 newtons-second-law
   relating parameters
     * force is a real
     * mass is a real
     * acceleration is a real

 world
   variables
     I, R, U is a real
     F, m, a is a real

 relation
   r1: ohms-law
     requiring arguments
       * I
       * R
     returning argument
       * U

   r2: newtons-second-law
     relating arguments
       * F
       * m
       * a

The relations between the parameters of ohms-law are directed and follow a clear input-output pattern. The relation between the parameters of newtons-second-law are undirected. That is, the requiring keyword indicates that all parameters that follow are inputs to the relation. The returning keyword indicates that all parameters that follow are outputs of the relation. The relating keyword indicates that one has not decided whether the parameters that follow are input or output of the relation.

All parameters within relation definitions are assigned a type. The type of each parameter must match the type of the respective positional argument upon instantiation. For example, the type of argument a must be equal to the type of parameter acceleration. This mechanism allows for automated type checking. That is, the compiler can automatically check if all relations are instantiated with the correct number and correct type of arguments.

So far, the proposed syntax requires the user to define all relation parameters a priori. However, many mathematical formulas can take an arbitrary number of arguments. For example, a summation or balance equation. Therefore, the one or more keywords are introduced which indicate that a relation requires one or more parameters of a certain type. For example, in the following listing the relation cost-summation requires one or more cost arguments as input and returns the variable total-cost as argument.

1
2
3
4
5
6
define relation
  cost-summation
    requiring parameters
      * one or more cost is a real
    returning parameter
      * total-cost is a real

Comments

ESL has two kinds of comments: annotation comments and code comments. Annotation comments are part of the system specification and are attached as documentation to components, variables, needs, goal, transformation, design, and relation specifications. That is, annotation comments can be used to specify additional arbitrary textual information that is not captured by the ESL concepts themselves. Code comments are not part of the specification but are used to write down information that is relevant to the engineers who are creating and reading the specification.


  1. N. Crilly. Function propagation through nested systems. Design Studies, 34(2):216–242, 2013. 

  2. G. Pahl and W. Beitz. Engineering design: a systematic approach. Springer Science and Business Media, London, England, 2007. 

  3. Robert B Stone and Kristin L Wood. Development of a functional basis for design. Journal of Mechanical design, 122(4):359–370, 2000. 

  4. A. H. Tilstra, C. C. Seepersad, and K. L. Wood. A high-definition design structure matrix (HDDSM) for the quantitative assessment of product architecture. Journal of Engineering Design, 23(10-11):767–789, 2012. 

  5. J. Hirtz, R. B. Stone, D. A. McAdams, S. Szykman, and K. L. Wood. A functional basis for engineering design: reconciling and evolving previous efforts. Research in engineering Design, 13(2):65–82, 2002. 

  6. J. Jiao and C.-H. Chen. Customer Requirement Management in Product Development: A Review of Research Issues. Concurrent Engineering, 14(3):173–185, 2006. 

  7. T. Wilschut, L. F. P. Etman, J. E. Rooda, and J. A. Vogel. Automated generation of a function-component-parameter multi-domain matrix from textual function specifications. Research in Engineering Design, 29(4):531––546, 2018. 

  8. Y. Umeda, M. Ishii, M. Yoshioka, Y. Shimomura, and T. Tomiyama. Supporting conceptual design based on the function-behavior-state modeler. Artificial Intelligence for Engineering Design, Analysis and Manufacturing, 10(4):275–288, 1996. 

  9. A. K. Goel, S. R. Rugaber, and S. Vattam. Structure, behavior, and function of complex systems: The structure, behavior, and function modeling language. Artificial Intelligence for Engineering Design, Analysis and Manufacturing, 23(Special Issue 01):23–35, 2009. 

  10. H. Komoto and T. Tomiyama. A framework for computer-aided conceptual design and its application to system architecting of mechatronics products. Computer-Aided Design, 44(10):931–946, 2012. 

  11. E. Hull, K. Jackson, and J. Dick. Requirements engineering. Springer, London, England, 2nd edition, 2017. 

  12. C. Arora, M. Sabetzadeh, L. C. Briand, and F. Zimmer. Requirement boilerplates: transition from manually-enforced to automatically-verifiable natural language patterns. In Proceedings of the 4th IEEE International Workshop on Requirements Patterns, 1–8. Karlskrona, Sweden, 2014.