gocr ep0242131-001
(PICTURE)(PICTURE) (PICTURE)Europäi _hes Patentamt _ (PICTURE) EuroPean Patent offi_e (PICTURE) Offi_e euroPeen des brevets _ PubIication number . O 242 131 B1 _ EUROPEA_ PATE_T SPECIFICATlO_ _ Date oF pubI_i_t_ion oF patent spec_ircat_ion .. _ _nt. __.6 .. _O6F 9J44, _O6F __/5O, 29.O3.9_ BUlletin 9_l_3 _o6F 3/o33, _o6F 3/o_3 _ AppIication number . 8_3o3_22.3 _ Date oF FiIing . oa.o4.8_ (PICTURE) (PICTURE)_ Graphi_al system ror modelling a pro_ess and asso_iated method. _ Note . Within nine months from the publication of the mention of the grant of the European patent, any penon may give notice to the European Patent Office of opposition to the European patent granted. _ Notice of opposition shall be fled in a written reasoned statement. It shall not be deemed to have been W rIed untiI the opposition Fee has been paid (A_. 99(1) European patent convention). (PICTURE)Jouve, 18, rue Saint-Denis, 75OO1 PARIS
gocr ep0242131-002
EP O 242 _3_ B_ Des_ription (PICTURE)BACKGROUND OF THE INVENTlON 5 (PICTURE)1 . Field of the Invention The present invention relates to systems for modeling processes and more particularly to computer sys- tems for modeling processes. 1o 2(PICTURE). Description of the Related Art Currently there is a strong movement toward very high level programming languages which can enhance programmer productivity by making a programming task more understandable and intuitive. The increasing use of computers by users who are not highly trained in computer programming techniques has lead to a sit- 15 uation in which the user's programming skills and ability to interact with a computer system often become a limiting factor in the achievement of optimal utilization of the computer system. There are numerous subtle complexities which a useroften must master before she can efficiently program a computer system. For example, typical earlier computer systems generally comprise software subsystems which include multiple programs, and such programs often utilize one or more subroutines. Software systems 2o typically coordinate activity between multiple programs, and each program typically coordinates activity be- tween multiple subroutines. However, techniques forcoordinating multiple programs generally differfrom tech- niques for coordinating multiple subroutines. Furthermore, since programs ordinarily can stand alone while subroutines usally cannot, techniques for linking programs to a software system generally differ from techni- ques for linking subroutines to a program. Complexities such as these often make it diffcult for a user, who 25 although she may be a specialist in her field is not a specialist in the computer field, to efficiently make use of powerful computer systems which are available for her use. The task of programming a computer system to model a process often is further complicated by the fact that a sequence of mathematical formulas, mathematical steps or other procedures customarily used to con- ceptually model such a process often does not closely correspond to the traditional programming techniques 3o used to program a computer system to model such a process. For example, a user of a computer system fre- quently develops a conceptual model for a physical system which can be partitioned into functional blocks, each of which corresponds to actual systems or subsystems. Computer systems, however, ordinarily do not actually compute in accordance with such conceptualized functional blocks. Instead, they often utilize calls to various subroutines and retrievals ofdata from different memory storage locations to implement a procedure 35 which could be conceptualized by a user in terms of a functional block. Thus, a user often must substantially master different skills in order to both conceptually model a system and then to cause a computer system to model that system. Since a user often is not fully proficient in techniques for causing a computer system to implement her model, the efficiency with which the computer system can be utilized to perform such modelling often is reduced. 4O One particular field in which computer systems are employed to model physical systems is the field of in- strumentation. An instrument typically collects information from an environment. Some of the types of infor- mation which might be collected by respective instruments, for example, include. voltage, distance, velocity, pressure, frequency of oscillation, humidity or temperature. An instrumentation system ordinarily controls its constituent instruments from which it acquires data which it analyzes, stores and presents to a user of the 45 system. Computer control of instrumentation has become increasingly desirable in view ofthe increasing com- plexity and variety of instruments available for use. In recent years, increasing effort has been directed toward providing more efficient means for implement- ing instrumentation systems. The task has been complicated by the fact that such systems include arbitrary combinations of hardware instruments and software components. The need for more efficient means of im- 5O plementation has been prompted by increasing demands for automated instrumentation systems and an in- creasing variety of hardware and software combinations in use. In the past, many instrumentation systems comprised individual instruments physically interconnected. Each instrument typically included a physical front panel with its own peculiarcombination of indicators, knobs, or switches. A user generally had to understand and manipulate individual controls for each instrument and 55 record readings from an array of indicators. Acquisition and analysis of data in such instrumentation systems was tedious and error prone. An incremental improvement in user interface was made with the introduction of centralized control panels. In these improved systems, individual instruments were wired to a control panel and the individual knobs, indicators or switches of each front panel were either preset or were selected to be 2
gocr ep0242131-003
EP O 242 _3_ B_ presented on a common front panel. Another significant advance occurred with the introduction of computers to provide more flexible means for interfacing instruments with a user. In such computerized instrumentation systems the user interacted with a software program of the computer system through a terminal rather than through a manually operated front 5 panel. These earlier improved instrumentation systems provided significant performance efficiencies over earlier systems for linking and controlling test instruments. Additional problems soon developed, however. Computer programs used to control such improved instru- mentation systems had to be written in conventional textual programming language such as, for example, ma- chine code, FORTRAN, BASIC, Pascal, or ATLAS. Traditional users of instrumentation systems, however, of- 1o ten were not highly trained in programming techniques and, therefore, implementation of such systems fre- quently required the involvement of a programmer to write software for control and analysis of instrumention data. Thus, development and maintenance of the software elements in these instrumentation systems often proved to be difficult. Some reasons for the difficulties associated with earlier computerized instrumentation systems included, 15 for example. (1) textual programming languages were non-intuitive and unfamiliar to the instrumentation sys- tem user; (2) traditional programming languages did not readily support the parallel activity of multiple individ- ual instruments; (3) concepts embodied in a computer program often were significantly different from con- cepts embodied in an instrumentation system's instrument hardware; (4) computer program software modules often did not match an instrumentation system's hardware modularity making interchangeability of software 2o and hardware difficult; and (5) techniques for designing, constructing, and modifying computer software were significantly different from corresponding techniques for developing an instrument hardware system. A computer based test system is disclosed in Electronics Design, Vol. 34, No. 8, 3 April 1 986, pp 149-1 52, 1 54, 1 56, New York, USA, which provides a common hardware interface between instrument modules, and a common programming environment through windowing software using an expanded version of Basic. A test 25 engineer specifies textually detailed steps of an operation in a first window, then uses a graphical represen- tation ofan instrument function in anotherwindow to produce code to drive the equipment to perform a desired step, and imports the code into the specification by indicating the step in the first window to which it should be assigned. A general type of program design technique involves data flow diagramming. Data flow diagramming typ- 3o ically involves an ordering of operations which is not specifically specified by a user but which is implied by data interdependencies. An advantage ofdata flowdiagramming is that it introduces parallelism into a computer system which, ofcourse, usually increases the speed and efficiency ofthe system. However as a design tech- nique, data flow diagramming has had difficulty representing conditional or iterative functions. Thus, designing using data flow techniques has been difficult. 35 Current attempts to develop text based data flow software systems generally have been unsuccessful. This lack of success may be due to difficulties in perceiving parallels between text based software and actual data flow, and it may be due to the fact that the text form of data flow generally suffered from the same diffi- culties in representing condition and iterative functions as did the diagram form of data flow. Thus, there exists a need for a system which can be relatively easily programmed for use in modelling a 4O process. Furthermore, there exists a need for an instrumentation system utilizing such a system. There exists a need for such a system which employs data flow techniques and overcomes the difficulties in representing conditional and iterative functions. The present invention meets these needs. Finally, there exists a need for such a system which represents data flow graphically rather than textually. 45 (PICTURE)SU__ARy OF THE lNVENTlON The present invention as claimed provides a method for modelling a physical process. A process typically can be characterized by one or more input variables and one or more output variables. The system includes a computer. It also includes an editor for displaying at least one diagram and for constructing execution instruc- 5O tions. The diagram graphically displays a procedure by which the one or more input variables can produce the one or more output variables. It also results in the construction of execution instructions which characterize an execution procedure which substantially corresponds to the displayed procedure. The system also includes an execution subsystem for assigning respective values for the one or more input variables and for executing the execution instructions to produce respective values for the one or more output variables. 55 These and other features and advantages of the present invention will become more apparent from the following detailed description ofexemplary embodiments thereof, as illustrated in the accompanying drawings. 3
gocr ep0242131-004
EP O 242 _3_ B_ (PICTURE)BRIEF DESCRIPTlON OF THE DRAWINGS The purpose and advantages of the present invention will be apparent to those skilled in the art from the following detailed description in conjunction with the appended drawings in which. 5 Fig. 1 shows a block diagram of a first system for modelling a process in accordance with the present in- vention; Fig. 2 shows a block diagram of a second system including the first system of Fig. 1 ; Fig. 3 is an illustrative drawing ofa representation ofa virtual instrument produced using the second sys- tem of Fig. 2; 1o Fig. 4 shows a block diagram of an instrumentation system including the second system of Fig. 2; Fig. 5 shows an illustrative front panel produced using the front panel editorofthe instrumentation system of Fig. 4; Fig. 6 shows an illustrative icon produced using the icon editor of the instrumentation system of Fig. 4; Fig. 7 shows an illustrative representation of an icon library which can be stored in a memory and used 15 to construct a block diagram using the block diagram editor of the instrumentation system of Fig. 4; Fig. 8 shows a graphical representation of a sequence structure; Fig. 9 shows a graphical representation of an iterative loop structure; Fig. 1 O shows a graphical representation of a conditional structure; Fig. 11 shows a graphical representation of an indefinite loop structure; 2o Fig. 12 shows graphical representations ofshift registers on the respective iterative loop structure of Fig. 9 and indefinite loop structure of Fig. 11 ; Fig. 1 3 shows an illustrative block diagram generally corresponding to the graphical representation of a sequence structure shown in Fig. 8; Fig. 14 shows an illustrative block diagram generally corresponding to the graphical representation of an 25 iterative loop structure show in Fig. 9; Fig. 1 5 shows an illustrative block diagram generally corresponding to the graphical representation of a conditional structure shown in Fig. 1 O; Fig. 16 shows an illustrative block diagram generally corresponding to the graphical representation of an indefinite loop structure shown in Fig. 11 ; 3o Fig. 17 shows an illustrative block diagram generally corresponding to the graphical representation of an iterative loop structure including a shift register shown on the left in Fig. 12; Fig. 18 shows a block diagram representing an exemplary data flow system; Fig. 19a illustrates avirtual instrumentdata structure diagram used bythefirstsystem of Fig. 1 , the second system of Fig. 2 and the instrumentation system of Fig. 4; 35 Fig. 19b shows a legend applicable to the illustration of Fig. 19a; Figs. 2Oa-l illustrate computer terminal displays during each successive step in a construction of an ex- emplary block diagram using the block diagram editor of Figs. 2 or 4; Fig. 21 is a drawing of an illustrative hardware instrumentation system of the present invention; Fig. 22 is a drawing representing a block diagram according to the present invention as displayed on a 4O computer console to model the illustrative hardware system of Fig. 21 . Fig. 23a shows the PROTO icon; Fig. 23b illustrates the startup screen with empty front panel window; Fig. 24 illustrates the FILE menu (ltems in clearsurrounds can be used and stippled items do not yetwork); Fig. 25 illustrates the EDIT menu; 45 Fig. 26 illustrates the FORMAT menu; Fig. 27 illustrates the CONTROL menu from the front panel window; Fig. 28 illustrates the NUMERIC control dialog box; Fig. 29 is an illustration of the STRING control dialog box; Fig. 3O shows the GRAPHICAL dialog box; 5O Fig. 31 illustrates an OPEN DIAGRAM selection from the FILE menu ofthe active front panel window pens and makes active a block diagram window; Fig. 32 illustrates employ, active Block-Diagram window created by choosing OPEN DIAGRAM from the FILE menu. Fig. 33 illustrates FUNCTlON menu from Block-Diagram window; 55 Fig. 24 shows a graphical representation of a sequence structure; Fig. 35 shows a graphical representation of a for loop structure; Fig. 36 shows a graphical representation of a case selection structure; Fig. 37 shows a graphical presentation of an indefinite loop structure; 4
gocr ep0242131-005
EP O 242 _3_ B_ Fig. 38 shows a graphical representation of a shift register in loop; Fig. 39 illustrates a dialog box obtained by choosing SPECIAL from the FUNCTlONS menu; Fig. 4O illustrates a ''fo_' loop structure layed down in the active block diagram window by selecting a glyph from the dialog box shown in Fig. 39; 5 Fig. 41 illustrates selection regions on structure glyphs (Moving dashded box encircles a glyph when it is selected); Fig. 42 illustrates that the the left side shows the proper position for placement of the grabber tool for re- sizing the structure glyph (Click-dragging produces the result shown in the right side ofthe figure. Release the mouse button and the enlarged structure will remain in place); 1o Fig. 43 shows two enlarged ''for'' loop structures rearranged for the Fibonacci instrument; Fig. 44 shows the dialog box obtained by choosing ARITHMETIC from the FUNCTlON menu. Fig. 45 shows a block diagram window after addition function was chosen from dialog box show in Fig. 44; Fig. 46 shows a block diagram window with addition and multiplication functions placed inside loop struc- 15 tures_, Fig. 47 shows a dialog box obtained by choosing CURVE FIT from the FUNCTlONS menu. Fig. 48 shows a block diagram window with CURVE-FIT object in place; Fig. 49 illustrates loop structure with shift register installed; Fig. 5O is an illustration of the final positions of constant objects alongside the inputs to the shift register; 2o Fig. 51 shows a pointer control in the front panel window obtained from the dialog box displayed by choos- ing the NUMERIC option from the CONTROLS menu; Fig. 52 shows the front panel window with one pointer control and two indicator controls; Fig. 53 shows the front panel window with graphical display repositioned to the upper right-hand corner of the window; 25 Fig. 54 shows the block diagram for the Fibonacci instrument with components in position ready forwiring. Fig. 55 illustrates thata connection is established between the slide-pointer control and the iteration vari- able box of the loop structure using the wiring tool; Fig. 56 illustrates ''hot spots'' for available functions; Fig. 57 is an illustration of the wired Fibonacci virtual instrument block diagram; 3o Fig. 58 is an illustration of examples of data types and their block diagrams representations; Fig. 59 is an illustration ofthe ICON EDITOR dialog box obtained by selecting the ICON EDIT PANE option from the FORMAT menu in the front panel window; Fig. 6O is an illustration of the icon editor following selection of the ICON button; Fig. 61 is a drawing in the icon editor; 35 Fig. 62 is an illustration displaying the instrument icon in the front panel window; Fig. 63 shows a terminal connector pane displayed in the upper left-had corner of the active window by choosing CONNECTOR PANE option from the FORMAT menu; Fig. 64 is an illustration of available arrangements of connector terminals; Fig. 65a is an illustration of the front panel window with selected terminal in connector pane; 4O Fig. 65b is an illustration of the highlighted terminal and control to be connected; Fig. 65c is an illustration of the connection between the established terminal and control; Fig. 66 is an illustration of the dialog box warning of consequences of clearing associations between con- trols and terminals; Fig. 67 (left) is an illustration of the virtual instrument to invert value set by control and display result on 45 indicator_, Fig. 68 (right) is an illustration of a diagram for a reciprocal virtual instrument; Fig. 69 is an illustration of an icon for the reciprocal instrument; Fig. 7O is an illustration of an instrument for taking reciprocals of and then adding two numbers; Fig. 71 is an illustration of the dialog box for selecting instrument to drop into block diagram window; 5O Fig. 72 is an illustration of an user-created reciprocal virtual instrument icon dropped into block-diagram window; Fig. 73 is an illustration of an icon for user-created invert-and-add instrument; Fig. 74 is an example solution for adding two resistances in parallel using two low-level, user-created vir- tual instruments in a hierarchical virtual instrument; and 55 Fig. 75 is an illustration of an dialog box from INSTRUMENTS option in the FUNCTlONS menu displaying two user-created virtual instruments. _
gocr ep0242131-006
EP O 242 _3_ B_ (PICTURE)DESCRIPTlON OF THE PREFERRED EMBODIMENTS The present invention comprises a novel method for modelling a process. The following description is pre- sented to enable any person skilled in the art to make and use the invention, and is provided in the context of 5 particular applications and their requirements. Referring to the drawing of Fig. 1 , there is shown a generalized block diagram 2O of a first system for mod- elling a process in accordance with the present invention. The first system 2O includes a block diagram editor 22, an execution subsystem 24 and a control processor 26. In the preferred embodiment, the block diagram editor 22 and the execution subsystem 24 are constructed in software. 1o As will be explained more fully below the block diagram editor 22 can be used to construct and to display a graphical diagram which visually and pictorially displays a procedure by which a value for an input variable can produce a value for an output variable. The procedure together with the input variable and output variable comprise a model of a process. Furthermore, the block diagram editor 22 constructs execution instructions which characterize an execution procedure which substantially corresponds to the displayed procedure. The 15 execution subsystem 24 assigns at least one value to the inputvariable and executes the execution instructions to produce a value for the output variable. The control processor 26 implements the block diagram editor 22 and the execution subsystem 24 of the preferred embodiment. The illustrative drawing of Fig. 2 shows a second system 28 for modelling a process in accordance with the present invention. The second system 28 includes a respective block diagram editor 3O, and an execution 2o subsystem 32 which are substantially identical to the block diagram editor 22 and the execution subsystem 24 of the first system 2O. The second system 28 also includes an icon editor 34 and a front panel editor 36. The second system 28 also includes a control processor 38 which is substantially identical to that ofthe first system 2O. The second system 28 permits a user to construct a virtual instrument 4O such as that represented in gen- 25 eralized form in the illustrative drawings of Fig. 3. The virtual instrument 4O includes a front panel 42 which permits interactive use of the virtual instrument 4O by a user. As will be explained more fully below, the front panel permits graphical representation of input and output variables provided to the virtual instrument 4O. The virtual instrument 4O also includes an icon 44 which permits use of the virtual instrument 4O as a subunit in othervirtual instruments (not shown). The virtual instrument 4O also includes a block diagram 46 which graph- 3o ically provides a visual representation ofa procedure by which a specified value for an input variable displayed in the front panel 42 can produce a corresponding value for an output variable in the front panel 42. The virtual instrument 4O itself is a hierarchical construction comprising within its block diagram 46 respective icons 48 and 5O referencing other virtual instruments indicated generally by respective blocks 52 and 54. The generalized block diagram of Fig. 4 shows an instrumentation system 56 incorporating the second 35 system 28 shown in Fig. 2. Elements ofthe instrumentation system 56 which are substantially identical to those ofthe second system 28 are referenced by primed reference numerals identical to those ofthe second system 28. The instrumentation system 56 includes a keyboard and display 58 and an instrument 6O. In a presently preferred embodiment, the control processor 38' and the keyboard and display 58 are implemented using a Macintosh Plus Computer manufactured by the Apple Computer Company of Cupertino , California. However, 4O a Macintosh Computer also produced by the Apple Computer Company could be used instead. The instrumentation system 56 can be used to control the instrument 6O, to acquire data from the instru- ment 6O to analyze that data, to store that data and to present that data. Figure 5 shows details ofan illustrative front panel 62 which can be produced using the front panel editor 36' and which can be displayed using the keyboard and display 58. It will be appreciated that the illustration of Fig. 5 represents an actual graphical com- 45 puter-generated display ofan exemplary front panel for the instrument 6O. The graphical representation of Fig. 5 illustrates physical control dials and switches for providing variable input information and illustrates a coor- dinate plane type indicator for displaying variable output information. More particularly, Fig. 5 shows a circular turn-dial and a slide switch for setting input variable data. The turn-dial and slide switch each correspond to respective rectangular boxes fordigitally illustrating variable input data in digital form. The illustrative front pan- 5O el also includes a coordinate plane display for illustrating variable output data. The graphical representations of input controls and output indicators are stored in a memory library, and a user may select from among a variety ofdifferentgraphical representations of input controls and output indicators in orderto construct a panel display which conforms to a user's intuitive understanding of how the instrument 6O is controlled and how it provides data. 55 Figure 6 illustrates an icon 64 which can be used to reference a front panel (not shown). A visual repre- sentation of the icon 64 can be produced using the icon editor 34'. The icon 64 corresponds to a particular front panel (not shown). As will be explained more fully below, the icon 64 can be used as a building-block in a hierarchial system constructed using the block diagram editor 3O'. The dashed lines of Fig. 6 indicate the 6
gocr ep0242131-007
EP O 242 _3_ B_ one-to-one correspondance between the icon 62 and the respective two-dimensional regions (or hot spots) 66 and 68 which correspond to respective variable input data and variable output data illustrated by controls and displays of the corresponding front panel (not shown). For example, the front panel might include input data in the form a sequence of samples and might provide output data in the form of an indicator showing vol- 5 tage reading per sample. The icon 64 then might be divided into two two-dimensional regions 68 and 66 which respectively correspond to the input sample count and the voltage reading for that sample count. Figure 7 illustrates a memory library including a plurality ofsoftware modules each representing a respec- tive front panel and each corresponding to a respective illustrative icon which references its respective front panel. As will be explained more fully below, a user may utilize the icons of the memory library to build a block 1o diagram; or helshe may build hislherown respective front panels and corresponding icons using the front panel editor 36' or the icon editor 34' in order to build a block diagram; or helshe may use both the icons of the icon memory library and icons which helshe produces himselflherselfusing the front panel editor 36' and icon editor 34' in order to build a block diagram. The drawings of Figs. 8-1 2 show the graphical representations ofstructures utilized in constructing a block 15 diagram as described below using the block diagram editor 3O'. The structures represented in Figs. 8-12 sub- stantially facilitate the application of data flow programming techniques which are used in the preferred em- bodiments of the present invention. Figure 8 illustrates a sequence structure. Figure 9 illustrates an iterative Ioop structure. Figure 1 O illustrates a conditional structure. Figure 11 illustrates an indefinite loop structure. Figure 1 2 illustrates a shift register on an iterative loop structure (on the left side) and a shift register on an 2o indefinite loop structure (on the right side). It will be appreciated that the graphical representations of the structures illustrated in Figs. 8-12 can be stored in a memory library as can execution instructions corresponding to the respective structures. Thus, a user can call upon a graphical structure library in order to display any one or more of the structures using the display facilities of the control processor 38' and keyboard and display 58 of the instrumentation system of 25 Fig. 4. The sequence structure, which has its graphical representation illustrated in Fig. 8, serves to divide a data- flow diagram into two subdiagrams, one representing an inside and another representing an outside of the sequence structure borders. The outside diagram behaves exactly as if the sequence structure and its con- tents were replaced by an icon with a terminal (or hotspot) foreach line crossing the sequence structure border. 3o The drawing of Fig. 8 shows a three-diagram sequence. In orderto minimize space used on a computerconsole screen, only one diagram of the sequence structure is visible at a time. Inside the structure border, multiple diagrams (not shown) can be constructed which execute in sequence. The sequence ofdiagrams are indicated by the respective numbers in the respective sequence diagrams. When the first diagram (indicated by the num- ber O) in this sequence completes its execution, the next one begins. The process is repeated until all diagrams 35 in the sequence have been executed. Each diagram in the sequence uses a subset of incoming signal paths and produces a subset of outgoing signal paths (the outgoing subsets must be mutually exclusive, but the incoming subsets are arbitrary). Con- stants may be used with any ofthe diagrams without any constraints. Variables used within a diagram are strict- Iy local to the sequence structure and may be assigned only once. Variables can be used multiple times in the 4O diagrams following the diagram where the variable was assigned. In accordance with data-flow principles, used in the preferred embodiments of the present invention, the sequence structure does not begin execution until all incoming signal paths have data available, and none of the outgoing signal paths produce data until all diagrams have completed execution. Figure 1 3 shows an illustrative block diagram 7O of a sequence structure. The sequence structure is cou- 45 pled to receive input signals on respective lines 72 and 74 and to provide respective output signals on respec- tive lines 76, 78 and 8O. Input registers 82 and 84 are provided to collect input data. A decoder 86 is provided to determine which computation andlor control element 88, 9O, or 92 to select, and a sequence counter 94 is included to undertake a count forsequencing between respective elements 88, 9O and 92. When all data inputs are present, an output of AND gate 96 becomes TRUE. This starts computation in computation andlor control 5O element 88 (assuming that it is the first element selected). When the control element 88 has completed com- putation, its output is stored in register 98. When the first element 88 has completed computation, the sequence counter 94 is free to advance by one. The decoder 86 will select the second computation element 9O. The output of AND gate 96 will become TRUE again and, computation will begin in the second element 9O. The output of the second element 9O will be stored in output register 1 OO. The sequence repeats for the third element 92, 55 and its output is stored in output register 1 O2. After the completion of the computation by the third element 92, the output data from all computations will be available forfurthercomputation by other instruments (not shown) of a block diagram system as will be explained more fully below. The iterative loop structure, a graphical representation of which is shown in Fig. 9, is similar to the se- l
gocr ep0242131-008
EP O 242 _3_ B_ quence structure in that the iterative loop structure partitions the data-flow graph into two parts. The intefiof diagram contains the body of the loop. Signal paths crossing the border of an iteration loop structure typically have a transformation applied. Incoming data are indexed in the most significant dimension so that the data inside the structure have dimensionality one less than outside. Outgoing data has the inverse transformation 5 performed. All signal paths crossing the border must have the same size for the indexed dimension. It is possible to disable the indexing on a signal path, in which case, the data behaves as if it were a constant available to each iteration. If indexing is disabled on an outgoing signal path, the data value is repeatedly overwritten and only the last value propagates out from the iteration structure. 1o There are two special variables which behave as constants within the body of the iterative loop structure. the numberof iterations, N, and the iteration numberor index, i. Usually, the numberof iterations to be executed is automatically set by the size of the dimension being indexed for an incoming signal path. In the event that there are no incoming signal paths, a scaler value must be specifically connected to the variable to specify the number of iterations. The iteration number is similar to a constant within the diagram except that its value 15 is O for the first iteration and increments by 1 at the end of each iteration. Iterations are independent and can, in principle, be executed in any order or completely in parallel except in the case where a non-reentrant virtual instrument (for example, a virtual instrument which cannot exeCute in parallel with itself such as one which manipulates or communicates with a physical instrument) is used. In that case, the iterations would be executed strictly sequentially. In accordance with data flow principles, all _o inputs must be available to start execution of an iteration loop. Furthermore, all outputs are generated after execution completes. Referring to the illustrative drawings of Fig. 14, there is shown a block diagram 1 O4 for an iterative loop. An iterative loop structure operates on data in an array one element at a time. The data for each element are sequentially stored in respective input buffers 1 O6 and 1 O8. A counter 11 O begins its count at O. When the first _5 data elements are available for both inputs of both respective input buffers 1 O6 and 1 O8, computation andlor control element 1 12 will generate outputs to be stored in respective output buffers 114 and 116. At that time, the counter 11 O will advance to 1 , and the process will repeat for the second data element in the array. This process will repeat until the counter 11 O reaches N-1 making a total of N computations. At that time a complete cycle signal will be generated by the comparator 1 1 8. The output signals stored in the respective output buffers 3o 1 14 and 116 then will be available for use by other computation instruments (not shown). The conditional structure, a graphical representation ofwhich is shown in Fig. 1 O, is similar in appearance to the sequence structure in its use of screen space, but it differs in its handling of signal paths crossing its border in that in each case a diagram may use any subset of incoming signal paths, but must produce all out- going signal paths. In accordance with data-flow principles, all inputs must be available in order to start exe- 35 cution. Furthermore, all outputs are generated after execution is completed. There must be a signal path that terminates at the case-selection terminal on the structure border. In the simplest case, a boolean-valued scalar is connected to the selector to select between case O (FALSE) and case 1 (TRUE). In the general case, a scalar number is connected to the selector. A special diagram (default case) may be specified to be used in the event that the selector does not lie in the range of specified cases. 4o The drawings of Fig. 1 5 illustrate a block diagram 12O corresponding to a conditional structure. The block diagram 1 2O forthe conditional structure is substantially similarto that ofthe block diagram 7O forthe sequenCe structure. The block diagram 12O for the conditional structure includes respective input registers 1 22 and 124, a decoder 1 26, an AND gate 128, three respective computation andlor control elements 1 3O, 1 32 and 1 34 and three respective output registers 1 36, 1 38, and 14O all coupled as shown in the drawings of Fig. 1 5. In operation, 45 the conditional structure block diagram 12O operates in a manner substantially similar to that of the sequence structure block diagram 7O, except that the decoder 126 of block diagram 12O is directly controlled by the case selection input provided on line 142 to select only one diagram. The indefinite loop structure, a graphical representation of which is shown in Fig. 1 1 , is similar in concept to the iterative loop structure in that the interior of the structure diagram represents the body of the loop, but 5o it differs in that signal paths crossing the border ofthe indefinite loop structure do not usually have an indexing transformation applied. There are two special variables applied within the body of the indefinite loop structure. iteration number or index, i, and recirculation flag, Q. The iteration number starts at zero and increments by one at the end of each iteration. A boolean value or expression is connected to the recirculation flag. A value of TRUE means 55 thatanother iteration will be performed. Ifthe recirculation flag is left unconnected, it is equivalent to connecting a TRUE constant. In accordance with data-flow principles applied in the preferred embodiments, all inputs must be available in order to start execution. Furthermore, outputs are generated after execution is complete. The illustrative drawings of Fig. 1 6 shows a block diagram 144 which corresponds to the graphical repre- 8
gocr ep0242131-009
EP O 242 _3_ B_ sentation of an indefinite loop structure shown in Fig. 1 1 . In operation, when data inputs are available on both respective input registers 146 and 148, an output of AND gate 1 5O will become TRUE to enable computation andlor control element 1 52. After computation is complete, output data are stored in respective output registers 1 54 and 1 56. After completion of the first loop, counter 1 58 increments, and the cycle begins again. This proc- 5 ess continues until a recirculation flag provided on line 16O goes FALSE. The output data are present after each cycle. A special construction available for use only within the respective loop structures is the shift register. A graphical representation of each respective loop structure type incorporating a shift register is show in Fig. 1 2. The shift registereliminates the need forcycles in a data-flowgraph making the result easierto comprehend 1o and to prove correct. The shift register behaves as if it were an ordered set of two or more variables, all of the same type and dimensionality. The first variable in a set is an output of the loop-body diagram and is located on the right border of the Ioop structure. The other variables of the set are inputs to the loop-body diagram and are located on the left border of the structure at the same elevation. 15 At the conclusion of each loop iteration, the data from the shift register output variable are shifted into the first input variable, and the previous value of the first input variable is shifted into the second input variable. The drawing of Fig. 1 7 shows an illustrative block diagram 162 illustrating operation of an iterative loop structure including a shift register. Respective latches 164 and 166 are provided to implement the shift register. In operation, the block diagram 162 of Fig. 1 7 (which represents an iterative loop structure with a shift register) 2o operates similarly to the block diagram 1 O4 of Fig. 14 (which represents an iterative loop structure minus a shift register) except that computation inputs are provided which give the system feedback from a previous cycle. An output provided by loop counter 168 is sensed by the comparator 1 7O. For the first loop, the multiplexer control 1 72 selects preselect inputs from respective preset gates 1 74 and 1 76. For all other cycles, respective 25 latches 164 and 166 are selected. The selected input is fed into the computation andlor control element 1 78. Data from input buffer 1 8O also is fed into the computation element 1 78. After each cycle, the computed output data are fed into respective output buffers 1 82 and 1 84. When the comparator 1 7O reaches N-1 , the process is completed, and the output data can be passed to a next instrument (not shown). The following discussion regarding data flow principles and a virtual instrument data structure diagram 3o will assist in understanding the operation of the respective block diagram editors 22, 3O and 3O' and the re- spective execution subsystems 24, 32 and 32' of the respective first system 2O, second system 28 and the instrumentation system 56. Referring now to the illustrative drawings of Fig. 1 8, there is shown a block diagram 1 86 of an exemplary data flow system. The block diagram 1 86 includes three respective input registers 1 88, 1 9O and 1 92 which 35 provide an accumulation of input data. As soon as all input data are present, an output of AND gate 1 94 will become TRUE, and computation andlor control element 1 96 will begin computation. The computation element 1 96 will begin generating output data which are stored in respective output registers 1 98, 2OO and 2O2. When all output data are available, an output token will be generated by the computation element 1 96 indicating that output data are available for transmission to a next system (not shown). It will be appreciated that the compu- 4O tation element can be a combination of more than one subsystem (not shown). Figure 1 9a illustrates a virtual instrument data structure diagram. The first system 2O of Fig. 1 , the second system 28 of Fig. 2, and the instrumentation system 56 of Fig. 4 each utilize the principles set forth in the data structure diagram of Fig. 1 9a. It will be appreciated that implementation of a system utilizing a data structure such as that diagrammed in the diagram of Fig. 1 9a advantageously permits the implementation ofan extended 45 data flow system like that illustrated in Fig. 1 8. Furthermore, it will be appreciated that implementation of the data structure like that of the diagram of Fig. 1 9a advantageously permits the implementation of a system in which execution instructions can be con- structed in a graphical fashion. More particularly, execution instructions can be constructed by constructing a visual display in which at least one input variable produces at least one output variable according to a dis- 5O played procedure. Furthermore, the execution instructions are constructed such that, when a value is assigned to a particular input variable, a value for a corresponding output variable is produced substantially according to the procedure illustrated in the visual display. Additionally, the execution instructions can be constructed in response to the construction of a block diagram comprising the graphical display. Thus, a user need only construct an appropriate visual display in order to construct the execution instructions. 55 Moreover, implementation of data flow principles by using a data structure such as that shown in the di- agram of Fig. 1 9a advantageously permit the use of parallel processing which increases the speed with which the execution of execution instructions can be accomplished. More particularly, Fig. 1 9a shows a system representation of a virtual instrument. Boxes 8a-8k, indicate 9
gocr ep0242131-010
EP O 242 _3_ B_ conceptual objects in the system that havewell-defined properties. Objects 8i, 8j and 8kare grouped intosahp- ed box 8s and share some properties and form a class of objects. As indicated in Fig. 13b which represents a legend applicable to the illustration of Fig. 19a, a solid line with an arrow is used to indicate a potential one-to- one relationship, i.e., the source object contains zero or 5 more destination objects (e.g., a vehicle containing zero or more wheels). Awiggly line with an arrow is used to indicate a potential one-to-many relationship; i.e., the source object may reference zero or one destination object (e.g., a library book may or may not have a borrower). Line 8n indicates that a virtual instrument 8b contains a front panel with a multiplicity of controls 8c. A control may be of clustered type in which case it contains a multiplicity of subcontrols as indicated by line 8p. 1o Line 8q indicates that a virtual instrument contains an icon with a multiplicity ofterminals 8d. Line 8l indicates that virtual instrument also contains a multiplicity of block diagrams 8e. In the system of the present invention, a virtual instrument either contains one diagram or none. Built-in virtual instruments representing primitive computations have no diagrams. Line 8r indicates that a block dia- gram contains a multiplicity of objects of the node class. A block diagram contains exactly one self reference 15 node 8i, and an arbitrary number of structure nodes 8j or instrument use nodes 8k. Line 8t indicates that a structure node contains a multiplicity of subdiagrams. Asequence structure or a conditional structure contains one or more sub-diagrams, and an iterative loop structure or indefinite loop structure contains exactly one subdiagram. Line 8m indicates that an instrument use node is used to reference another virtual instrument. The instrument use node may reference a virtual 2o instrument in real-time; or it may reference previous data acquired by the virtual instrument. Line 8u indicates that each object ofthe node class contains a multiplicity ofterminals 8g. Line 8v indicates that a block diagram also contains a multiplicity ofsignal paths 8f. Each signal path contains a multiplicity ofterminals as indicated by line 8w. There is at most one terminal per signal path that is designated as the source of the signal. Each terminal contained in a signal path also is contained in a node. However, there may be terminals in nodes which 25 are not in signal paths. The terminals in a signal path are typically in different nodes. Lines 8y and 8z indicate that each terminal may reference a front panel control or a block diagram control (e.g., a constant). Aterminal references exactly one control, and it is either on the front panel or on the block diagram. Figs. 2Oa-l illustrate computer screen displays during each successive step in a construction ofan exem- plary block diagram using a block diagram editor such as that of Figs. 2 or 4. (lt should be appreciated that 3o the block diagram editor of Fig. 1 operates in a similar fashion; except that it does not include an icon editor or a front panel editor for user-constructed front panel). More particularly, in Fig. 2Oa, a control knob is placed in the front panel, and its associated terminal au- tomatically appears in the block diagram. The system representation shows the virtual instrument with a dia- gram containing a self reference node, and a terminal in the self reference node which references the front 35 panel control. In Fig. 2Ob, a control graph indicator type is placed in the front panel, and its associated terminal auto- matically appears in the block diagram in the same position relative to the other terminal as the graph is to the knob. This makes it possible to distinguish the terminal even without supplementing the graphics with text lab- els. 4O In Fig. 2Oc, a constant with value 2O is placed in the block diagram, and is reflected in the system repre- sentation by another terminal and control attached to the self reference node. In Fig. 2Od, an icon referring to a built-in virtual instrument is placed in the block diagram. (An alternative view of the block diagram could show the icon terminals instead of the icon itsel_. The system representation shows another node of instrument use type in the virtual instrument diagram and three terminals and controls 45 corresponding to the terminals and controls the referenced virtual instrument. In Fig. 2Oe, an iterative loop structure is placed in the block diagram. The system representation shows the structure node in the diagram along with terminals and controlsforthe loopvariables. Note thatthe iteration number is accessible only from within the loop; while the iteration limit is available inside and outside the loop as evidenced by the two terminals which reference it, one in the structure node and the other in the self- 5O reference node of the diagram within the structure node. In Fig. 2Of, an icon referencing another built-in virtual instrument is placed inside the iterative loop struc- ture. In Fig. 2Og, a wire is connected from the terminal associated with the front panel knob to the loop limit terminal of the loop structure. The front panel knob terminal is determined to be the signal source. 55 In Fig. 2Oh, a wire is connected from the iteration number terminal to a terminal on the virtual instrument inside the loop. This signal path lies completely within the loop structure subdiagram. The system represen- tation shows the signal path with the iteration number terminal and the terminal on the instrument use node. The iteration number terminal is determined to be the signal source. _O
gocr ep0242131-011
EP O 242 _3_ B_ In Fig. 2Oi, the constant is wired to a terminal of the virtual instrument within the loop. In this case, the wire crosses the structure border so that a pair of terminals and a control are created, and the wire is split into two signal paths, one outside the loop structure and one inside. The constant is determined to be the source terminal of the outside signal, and the inside terminal at the border is determined to be the source of the inside 5 signal. In Fig. 2Oj, a wire is drawn from the virtual instrument inside the loop to the virtual instrument outside the Ioop. This wire crosses the border so it is split into two signal paths. The wire on the outside is thicker because it represents an array signal path (as will be explained more fully below). In Fig. 2Ok, the iteration number terminal is connected to a terminal on the outside virtual instrument. The 1o wire crosses the border so that it is split into two signal paths. The inside signal path is joined to the existing signal path originating at the iteration number terminal. Again, the wire on the outside is thicker. In Fig. 2O1 , the output of the virtual instrument outside the loop is connected to the terminal associated with the front panel graph. The wire pattern indicates that it represents a clustersignal path (as will be explained more fully below). 15 The following description describes the operation of the execution subsystems 24, 32 and 32' of the re- spective first system 2O, second system 28 and the instrumentation system 56. The first step in the execution of a virtual instrument is accomplished by executing its block diagram. The first step in the execution of a block diagram is accomplished by scanning the terminals contained in the dia- gram's self-reference node. For each terminal which is the source of a signal, the data token is moved from _o the control referenced by the terminal to the terminal itself. The second step in the execution of a diagram is to initialize the token short-count of each node in the diagram to the number of input signals to that node. The third step in the execution of a diagram is to propagate signals from the self-reference node. Propagation of signals from a node is accomplished by scanning all of the node's terminals. For each terminal that is source of a signal the data token on the terminal is copied to each destination terminal of the signal path. Each token _5 placed on a destination terminal causes the short-count ofthe node containing the terminal to be decremented. If it becomes zero in the process, then that node is scheduled to execute. The first step in the execution of a node is accomplished by copying the tokens from the node's terminals to the reference controls. The second step depends upon the type of node. For an instrument use node that references a real-time virtual instrument, the next execution step is to copy the tokens from the node's controls 3o to the virtual instrument's controls and to execute the virtual instrument. For an instrument use node that ref- erences previously stored data of a virtual instrument, the tokens from the appropriate data record are read in and placed on the node's terminals. For a sequence structure node, the next step is to execute the first sub- diagram. For a conditional structure node, the next step is to execute the subdiagram indicated by the value of the token on the selected control. For an iterative or indefinite loop structure node, the next step is to set 35 the value of the token on the iteration number control to zero and to execute the subdiagram. For a self- reference node, the next step is to perform the next step in the execution of the node or the virtual instrument which contains the diagram that contains the self-reference node. The third step in the execution of a node also depends upon the type of node. For an instrument use node or a conditional structure node the output data tokens are propagated along the signal paths. For a sequence 4o structure node, the next subdiagram is executed, if one exists, and if not, the output tokens are propagated. For a loop structure node, the shift registers are clocked (the data is shifted), the iteration numbers increment- ed, and the subdiagram is re-executed, if appropriate; otherwise the output tokens are propagated. The second step in the execution of the virtual instrument is to log the tokens on the front panel controls if data logging is enabled. The third step in the execution of the virtual instrument is to copy the tokens from 45 the virtual instrument's indicators to the instrument use node's output terminals and to schedule the instrument use node to execute its next step. The third step ofvirtual instrument execution is performed only if the virtual instrumentwas executed in response to an instrument use node request. Ifthe virtual instrumentwas executed interactively, there is no third step. Figure 21 shows an illustrative hardware configuration of an instrumentation system 2O4 according to the 5O present invention. The system 2O4 includes a Macintosh computer 2O6 with an interface to a GPIB (general purpose instrument bus) 2O7, a Tektronix 5O1 O Function generator 2O8, a Fluke 884OAdigital multimeter 21 O, and a unit undertest 212 all coupled as shown. The traditional approach to automating this measurementwould be to create a program in a computer language such as BASIC to interact with the test instruments 2O8 and 21 O to collect, format, and display in real time the measured values pertaining to the unit under test 212. 55 Figure 22 shows a drawing of a computer-generated display of a completed block diagram for the design example of Fig. 21 . This block diagram is the graphical program representing the instrument's operation. It shows the interconnections between the elements of the instrument, the signal paths, and the relationship to other virtual instruments. At the upper left ofthe di_a_gram, fourfront panel input controls are shown connected
gocr ep0242131-012
EP O 242 _3_ B_ to a ''Ramp'' icon. This icon is built-in function which takes input minimum and maximum values, number of steps, and a flag to indicate linear or log steps, and produces as output an array ofsample points, in this case the frequencies of interest. The output is a bold one, which indicates that the data type is an array. The large rectangular region in the center ofthe diagram is an iteration loop structure. The diagram placed 5 inside this region will be executed multiple times. for i = O to N-1 . The inputs to the iteration loop are amplitude and the array of frequencies, shown on the left. The amplitude remains the same data type as it crosses the Ioop boundary. The array, however, as shown by the change in line type, is automatically indexed as it crosses the boundary to provide one array element per iteration. Since there is no connection to the loop-count (the N at the upper left corner), the length of the input array is used by default as the number of iterations. 1o Inside the iteration loop are two virtual instrument icons. The first takes as input an amplitude and a fre- quency and performs the appropriate IEEE-488 operations to set the function generator 2O8 of Fig 21 . The second performs the appropriate IEEE-488 operations to obtain a voltage measurement from the multimeter 21 O of Fig. 21 . The dotted line indicates that there is no data flow, but ensures that they execute sequentially. These two icons represent simple virtual instruments that are easily designed using built-in high level IEEE- 15 488 functions to communicate with the multimeter 21 O. Each iteration ofthe iteration loop produces one voltage measurement. This results in an array ofvalues, as shown by the line change as it exits the loop at the right. The graph function takes this array and the array offrequencies as input and produces as output the data structure for the front panel graph indicator. Note the self-documenting effect of the graphical language, with the iteration loop structure contributing to the read- 2o ability of the program. With the front panel and block diagram complete, the instrument is ready to be used. The instrument is operated from the front panel. To execute the instrument, the user simply configures the input controls and ''clicks'' the GO button on the top of the screen (as will be appreciated from the description below). 25 (PICTURE)Operation of the Preferred Embodiments The presently preferred embodiments ofthe invention are implemented in software. Thefollowing descrip- tion explains the operation ofthe presently preferred embodiment as implemented using eitheran Apple Mac- intosh Plus Computer or an Apple Macintosh Computer. The explanation is intended to be illustrative and not 3o exhaustive and uses specific examples to explain the principles of operation of the preferred embodiments. Itwill be appreciated thatthe principles ofoperation explained belowwill serve as a basisforthe heuristic learn- ing process necessary to fully appreciate and practice the full scope of the present invention. Thus, it is not intended that the scope and content of protection afforded to the present invention as claimed be limited in any way by the following explanation. 35 The following explanation explains how to use the implemented functions of the preferred embodiments. A walk through of the display menus is provided first, with working functions described. This presentation is followed by a walk through of an example application program that computes Fibonacci numbers. To start, double click on PROTO icon. The PROTO icon is illustrated in Fig. 23a. The opening screen is shown in Fig. 23b. This screen is the front panel window on which controls can be 4O placed. The tools and controls below the menu bar will be described later. Figure 24 shows the contents of the FILE menu. In this and subsequent screen pictures, items in clear surrounds can be selected. Stippled items do not yet work. The NEW, OPEN, CLOSE, SAVE and SAVE AS... items work as in any other Maclntosh program that follows Apple's user-interface guidelines. NEW opens a new instrument, OPEN brings up a dialog box with existing instruments' names, CLOSE closes the current 45 instrument, SAVE saves the instrumentwith the current file name, and SAVE As... saves a copy ofthe current instrument under the user-specified instrument name. OPEN DIAGRAM opens the block diagram window. Its use will be describe later. QUIT exits from PROTO and returns the user to the Finder. Figure 25 shows the EDIT menu selections. This one is easy. Only CLEAR works. CLEAR is useful for removing items from the active window, e.g., selected wires and structures from the block diagram window, 5O or controls from the front panel window. Figure 26 shows the FORMAT menu. The FORMAT menu has five working items. The first two FULL SCREEN and TILE alter the screen display. The FULL SCREEN options produces over-lapping windows. The TILE option produces non-overlapping windows. Both the FULL SCREEN and TILE selections can be reverted by selecting them again. The TILE option is dimmed if a block diagram window is not open. 55 Figure 27 shows the contents of the CONTROL menu. The CONTROL menu is only available when the front panel window is active. The NUMERIC, STRING and GRAPHICAL items all workto some degree. Choos- ing the NUMERIC item brings up the dialog box shown in Figure 28. This box offers the user many choices, however, only a few are implemented at this time. The small glyphs half-way down the left side ofthe box offer _2
gocr ep0242131-013
EP O 242 _3_ B_ the choice of display style. clicking on the chosen glyphs produces a reverse-video display of that pictograph. MINIMUM, MAXIMUM and DEFAULT work by selecting their adjacent frames and filling in the desired values. Arrays can be specified by choosing the number of dimensions and inserting that number into the frame to the leftofARRAY DIMENSlONS. The numberofelements in each array dimension is then specified by entering 5 appropriate values into the dimension frames numbered 1 to 8. Selecting the INDICATOR ONLY box makes the control a passive indicatoron the front panel. Selecting SHOW DIGITAL READOUT provides a digital read- out of the value of the control. The current value of the control can be altered by adjusting the control itself or by altering the value in the digital read out by selecting and entering values from the keyboard. Note the CONTROL menu is dimmed when the block diagram window is the active window. It is made avail- 1o able by making the front panel window the active window. Figure 29 shows the dialog box for the STRING control. String controls can be filled with values for use as controls or indicators. This capability is useful for sending command or address strings for writing to and reading from a GPIB. Figure 3O shows the dialog box for the GRAPHICAL control. This control is used to display graphs. It must 15 be used with the INDICATOR ONLY box selected. The FUNCTlONS menu name is dimmed when the front panel window is active. It becomes available when the block diagram window is made the active window. Figure 31 shows the FILE menu displayed from the front panel window and the OPEN DIAGRAM item se- Iected. Selecting this item opens a block diagram window. Figure 32 shows the screen for a new block diagram _o window. Notice that the untitled block diagram window is distinguishable from an untitled front panel window by the presence of the black bullets (_____) in the title bar and the dimming of the CONTROL menu name. A block diagram is created by choosing items from the FUNCTlONS menu. Figure 33 shows the contents of the FunG tions menu. Only a subset offunctions from the items SPECIAL, ARITHMETIC, COMPARATIVE, INPUT-OUT- _5 PUT are working. Single functions are also available from STRING, SIGNAL PROCESS and CURVE FIT. The working functions from each of the menu options will be described in turn. The functions available from the SPECIAL option of FUNCTlONS menu include STRUCTURES, SPECIAL ITEMS, and CONSTANTS. Each of the available functions is described below. 3o (PICTURE)Structures To remedy shortcomings of the conventional data flow, which makes it substantially unusable as a pro- gramming language, the present invention includes graphical operators. These structures implement in a data flow environment what typical structured programming control do. 35 (PICTURE)Sequence Structure The sequence structure in its simplest form serves to divide the data flow diagram into two subdiagrams, the inside of the structure and the outside of the structure. The outside diagram behaves exactly as if the 4o sequence structure and its contents were replaced by an icon with hotspots foreach line crossing the structure border. Figure 34 shows a 3D view of a three diagram sequence; however, to minimize screen space, only one diagram at a time is visible. Inside the structure border are multiple diagrams which execute in sequence. This sequence is indicated by the numeral in the upper leftcorner. When the first diagram in the sequence completes 45 execution, the next one begins. The process is repeated until all the diagrams in the sequence have been exe- cuted. Each diagram in the sequence uses a subset of the incoming arcs and produces a subset of the outgoing arcs (the outgoing subsets must be mutually exclusive to avoid arc fan in, but the incoming subsets are arbi- trary). Constants may be used within any of the diagrams without any constraints. Variables used within a di- 5O agram are strictly local to the sequence structure and may be assigned only once. Variables can be used mul- tiple times in the diagrams following the diagram where the variable was defined. The sequence structure does not begin execution until all the incoming arcs have data available, and none of the outgoing arcs produce data until all the diagrams have completed execution. There may be at most one arc that terminates at the sequence structure border itself, acting as a trigger for the sequence (it is just one 55 more input which must be present for the sequence to begin). An arc originating at the sequence structure bor- der may be used as a ready signal indicating that the sequence has completed execution. _3
gocr ep0242131-014
EP O 242 _3_ B_ (PICTURE)lnteractive Loop (''for'' loop) Structure The iterative loop structure is similar to the sequence structure in that it partitions the data flow graphic into two pieces. The interior diagram represents the body of the loop and is similar to a sequence structure 5 with N identical diagrams (Fig. 35). Arcs crossing the borderofan iteration loop structure typically have a trans- formation applied. Incoming arcs are indexed in the most significant dimension so that the data inside the struG ture has dimensionality one less than outside. Outgoing arcs have the reverse transformation performed. All arcs crossing the border must have the same size for the indexed dimension. It is possible to disable the indexing in which case an incoming arc behaves as if it were a constant available to each iteration. If in- 1o dexing is disabled for an outgoing arc the value is repeatedly overwritten and only the last value propagates out from the iteration structure (this is one difference from the way N sequence diagrams would behave). By default, the most significant dimension is the one indexed but this may change to any dimension. There are two special variables that behave as constants within the body of the iteration loop structure. the number of iterations (N) and iteration number or index (i). 15 The number of iterations to be executed is automatically set by the size of the dimension being indexed. In the event that there are no incoming arcs a scalar value must be specifically connected to this variable to specify the number of iterations. The other variable is the iteration number (ranging from O to N-1 , as in C). Iterations are independent and could in principle be executed in any order or completely in parallel, except in the case where a non-reentrant virtual instrument is used. In this case, the iterations behave exactly like _o N identical sequence diagrams. As usual, the iteration loop does not begin execution until all the incoming arcs have data available, and no outgoing arcs have tokens generated until all iterations have completed execution. An arc originating at the iteration loop structure border may be used as a ready signal indicating that the loop has completed execution. _5 (PICTURE)Case (Conditional) Selection Structure The case (conditional) selection structure of Fig. 36 is similar to the sequence structure in its usage of screen real estate, but it differs in its handling of arcs crossing the border in that each case diagram must use all incoming arcs and generate all outgoing arcs. The usual execution rules apply (all inputs available to start, 3o all outputs generated simultaneously after all execution completes). There must be exactly one arc that terminates at the case selection structure border itself, acting as a selectorforthe case. In the simplest case a boolean valued scalar is connected to the selectorto select between case O (false) and case 1 (true). In the general case a scalar number (or string variable, if it is specified as a discrete list of strings) is connected to the selector. 35 A special diagram (default case) may be specified to be used in the event that the selector does not lie in the range of specified cases. An arc originating at the case selection structure border may be used a a ready signal indicating that the selected case has completed execution. Indefinite Loop (''do-while'' loop) Structure 4O (PICTURE) The indefinite loop structure is similar in concept to the iterative loop structure in that the interior diagram represents the body of the loop, but it differs in that arcs crossing the border of an indefinite loop structure typically do not have an indexing transformation applied. Figure 37 shows a graphical representation of an in- definite loop structure. 45 There are two special variables available within the body ofthe indefinite loop structure. One is the iteration number which starts with a value of O and the other is the recirculation flag to which the value of the ''while'' expression is connected (if left unconnected it is equivalent to ''while(true)''). Iterations are typically related so the operation of the indefinite loop is similar to an indefinite number of identical sequence diagrams. The usual execution rules apply. There may be at most one arc that terminates 5o at the indefinite loop structure border itself, acting as a triggerfor the loop (it is just one more input which must be present for the loop to begin). An arc originating at the indefinite loop structure border may be used as a ready signal indicating that the loop has completed execution. Special Items. Shift Registers 55 (PICTURE) There is one special construction which is available for use only within the loop structures. This construG tion is the shift register. The shift register eliminates the need for cycles in a data flow graph making the result much easier to understand and to prove correct. The shift register behaves as if it were an ordered set of two _4
gocr ep0242131-015
EP O 242 _3_ B_ or more variables all of the same type and dimensionality. The first one in the set must be a left node within the loop body diagram. The others must be root nodes within the loop body diagram. To include a shift register in the indefinite loop structure, one must first highlight the desired structure and then select the shift registerfrom the functions menu item. The indefinite structure 5 now shows two boxes on its left edge (root nodes) and one box on the right edge (leaf node). See Fig. 38. At the conclusion of each iteration of the indefinite loop all shift registers within the body are ''clocked'', i.e., the value of the leaf node is placed in the first root node while the previous value of the first root node is placed in the second root node, etc. Prior to the first iteration of the indefinite loop, root nodes of shift registers are set to false, zero, or empty 1o string, as appropriate. These default initial values may be overridden by specifying alternative constants, or by setting the nodes using incoming arcs from arbitrary external expressions. (PICTURE)Using PROTO. The Fibonacci Number Problem 15 The best way to learn to use the presently preferred embodiment of the invention is to create a working virtual instrument. The following exercise is a standard one from introductory textbooks on computer program- ming. Creating a Fibonacci sequence. In place of the traditional programming languages like C or Pascal, we will use the presently preferred embodiment of the invention. The Fibonacci sequence (1 , 1 , 2, 3, 5, 8, 13, 21 , 34 ....) arises in such diverse applications as the number _o of petals in a daisy, the maximum time it takes to recognize a sequence of characters, and the most pleasing proportions for a rectangle. It is defined by the relations. f(1) = 1 f(2) = 1 _5 f(n) = F(n - 1) + f(n - 2) for n > 2 Starting with the third term, each Fibonacci number is the sum ofthe two previous Fibonacci numbers. In the Fibonacci instrument described below, the presently preferred embodiment is used to create a virtual in- strument to generate and display a sequence of Fibonacci numbers. To accomplish this a number offeatures of the preferred embodiment are used including iterative-loop structure arithmetic functions, numeric and 3O gfaphIC COntfOlS and COnStantS. (PICTURE)Getting Started To get started double click on the PROTO icon You new screen should like Fig 1 Construction ofvirtual 35 ' ' ' ' inStfument Can Staft ffom eithef the Ffont-Panel of the bloCk diagfam windowS, howevef, SinCe labeIS fof the controls are not yet implemented, it is perhaps easier to begin from the block diagram window. To reach the block diagram window pull down the FILE menu and choose OPEN DIAGRAM. This action will open an active block diagram window (see Fig. 32). You are now ready to build your virtual instrument. Pull down the FUNCTlONS menus and choose SPECIAL This choice will produce a dialog box with a dis- 4O ' play offunCtionS (See Fig. 39). The fouffunCtionS fepfeSented on the fifSt line in the dialog box afe deSCfibed above. All fourofthese functions are implemented. An item is selected by placing the cursor on it and clicking. This action closes the dialog box and deposits the chosen item in your active block diagram window. Figure 4O shows an example of a ''for'' loop structure layed down in the block diagram window in this manner. Items chosen from the FUNCTlONS menu are always deposited nearthe centerofthe window and if pre- 45 ' viouS itemS afe not moved new itemS may obSCufe old itemS. Items may be selected and moved about in the activewindow by using the grabberglyph. Placing the cursor on the grabber glyph in the tool palette in the upper left portion of the screen just beneath the menu bar and clicking will change the cursor into the grabber when you move the mouse back down into the active-window 5o area. The TAB key can also be used to change the tool. Pressing the TAB selects the next cunor in the palette. Note. the ''hot spot'' on the grabber icon is at the base of the thumb and first finger. Several actions, including moving objects, require that the object be ''selected'' before some action can be performed. To select a structure its ''select regions'' must be identified. The ''for'' loop structure object has two select regions, one along the rightedge and a second along the bottom edge ofthe object. Figure 41 shows 55 the location ofthese regions. Positioning the grabber's hot spot in one ofthese regions and clicking will select the object. The case and ''do-while'' structures are selectable anywhere along their borders. The sequence structure is also selectable along its borderexceptfor the uppercorners which perform the function ofshifting forward and backward through the sequence. When selected, the object is surrounded by a moving dashed __
gocr ep0242131-016
EP O 242 _3_ B_ box. The object can now be moved by placing the grabber on the object and click-dragging to a new location. Releasing the mouse button will leave the object selected in its new location. Moving the cursor off the object and clicking will de-select the object. Sometimes you may need to duplicate the structure you are working with. You could return to the FUNC- 5 TlON menu, choose SPECIAL again and click on the object you want to use. There is a short-cut to duplicate objects from within the block diagram window. To duplicate the object from within the active window, use the grabber tool to select the objection to be duplicated, press the OPTlON key and click-drag a copy offthe orig- inal. Note. Select can be an ambiguous term with reference to the Mac user interface. It is not synonymous 1o with choose, rather, it is used to indicate that the object is to be ''selected'' by placing the cursor into a select region for that object and clicking the mouse button. Two ''for'' loop structures are required for the Fibonacci instrument we are building. Use one ofthese two methods to duplicate a copy of the structure in your active window. Your active window should now contain two identical ''for''loop structures. 15 It is often necessary to place objects within objects. To do this, it is helpful to be able to resize objects. The structures can all be resized in the same fashion. Place the grabber tool in the lower right-hand corner of the structure and click-drag the objectto the desired size. Figure 42 shows the proper placementofthe grabber for the resize operation and the results of enlarging the original structure object. Additional functions are required in both loop structures for our instrument. Resize both loop structures _o to be about the same size of the larger objection in Fig. 42. Now reposition the two structures into locations similar to those shown in Fig. 43. The Fibonacci virtual instrument uses two arithmetic functions. addition and multiplication. These func- tions are obtained by choosing ARITHMETIC from the FUNCTlONS menu (see Fig. 33). Choosing this menu item opens the dialog box shown in Fig. 44. The functions currently implemented are. addition, subtraction, _5 multiplication and division. Choose addition by clicking on the addition function glyph. Figure 45 shows the result of choosing the addition function. The dialog box is closed and an object is placed near the center of the Block-Diagram window. Select this object by placing the grabber on it and clicking. The entire area of the arithmetic-function object is a select region. The selected object can now be click-dragged to a new position. Move the addition-function object to a location on the upper loop structure similar to the position shown in 3o Fig. 46. Select a multiplication function and position it in the lower loop similar to Fig. 46. Three more functions are needed before we move to the front panel and install some controls and indica- tors. First, pull down the FUNCTlONS menu again and select CURVE FIT. This choice will display the dialog box shown in Fig. 47. At this time only the linear-fit function is implemented. Choose LINEAR FIT. The object will drop approximately in the center of the screen. Click-drag the object to a position similar to the one illu- 35 strated in Fig. 48. The next function we need is a shift register. This function must be used within a control structure, there- fore a structure object must be selected before the shift register is selected from the SPECIAL menu item. Selectthe top loop, pull down the FUNCTlON menu and select SPECIAL. From the SPECIALdialog box choose the shift register glyph from the row of glyphs labelled. SPECIAL ITEMS (see Fig. 39). Figure 49 shows what 4o the loop structure looks like with the shift register function installed. Operation of the shift register function was reviewed above. The last functions to be implemented in the block diagram window are constants. Constants are available from the FUNCTlONs menu from the menu item labelled SPECIAL. We need two constants. Choose a numeric constant glyph from the dialog box and duplicate its object in the window. Figure 5O shows the two constant 45 objects moved into position along-side the inputs to the shift register. Set one constant to 1 and the other to O by using the operator (the ''pointing finge_') tool and double clicking in the constant box then typing in values from the keyboard. Now we will move to the front panel window to implement some controls and indicators. Click on the ex- posed edge of the front panel (see Fig. 32). The Front-panel window should come to the front with an empty 5O window. Notice that now the FUNCTlONS option in the menu bar is dimmend, by the CONTROLS option is darkened indicating that it can be accessed. Pull down the CONTROLS menu and choose NUMERIC. The con- tents of the CONTROLS menu are described above. Choose the NUMERIC option from the menu. The dialog box should look like the one in Fig. 28. Move your cursor to the area ofthe three display styles on the left side of the dialog box and select the center glyph 55 representing a ''pointe_' control. Notice that then selected the image ofthe glyph will change to reverse video. Move the cursor to the box labelled MINIMUM and double click to select it. If the interior of the box does not darken move your cursor around inside the box while clicking until you find the right spot. When the interior is darken enter 1 from the keyboard. Move the cursor to the box labelled MAXIMUM and enter 2O. Now move _6
gocr ep0242131-017
EP O 242 _3_ B_ the cursor to the box labelled OK and click. This action will close the dialog box and return you to the front panel window where you should see a larger representation of the pointer control you selected from the dialog box inside a moving dashed rectangle. Controls can be moved about the front panel window just as objects were moved in the block diagram window. Notice that the behaviorofthe controls is more like the non-structure 5 objects in the block diagram window. the entire control is electable, there is not a limited select region like on the structures. Click-drag the pointer control to the upper left-hand region of the Window to a position similar to the one shown in Fig. 51 . This control will determine the length of the Fibonacci sequence that your instru- ment calculates. Return to the CONTROLS menu and again choose the NUMERIC option. This time select the box labelled 1o INDICATOR ONLY and choose the DIGITAL CONTROL glyph (the right-most control glyph), then click in the OK box. The front panel window should now display a rectangle within a rectangle with the interior rectangle containing a O. This control is a passive indicator. It will be used to display intermediate was well as final values ofour Fibonacci instrument. We need a second indicator display. Controls can be duplicated in the same man- ner as objects in the block diagram window. Use the grabber tool. First select the control to be duplicated, de- 15 press the OPTlON key and click-drag a copy off of the original. Arrange these two indicators to the left of the front panel window similar to the positions shown in Fig. 52. We need one more front panel control to complete or instrument, a graphical display. Return to the CON- TROLS menu one more time and choose the GRAPHICA1 option. This option will produce a dialog box like the one shown in Fig. 29. Choose INDICATOR ONLY. Do this and click the OK box to return to the front panel 2o window. There should now be a rectangular-shaped graph with linear coordinates setting in the middle of the window. This is the graphical display control. It can be moved just like the objects in the block diagram window and the other controls on the front panel. It can also be resized, as can the other controls. The procedure for resizing is identical to the procedure used to resize objects in the block diagram. select the control and then click-drag with the grabber tool from the lower right-hand corner of the selected area. The situation with the 25 NUMERIC controls is the same but because the fullness ofthe select region the area to ''grab'' with the grabber is less obvious, but it is the same region just interior to the lower right-hand corner of the select area. It may take a little practice to be able to reliably resize these controls. Move the graphical display to the upper right- hand corner of the window. Your screen should now look like the one shown in Fig. 53. At this point we can return to the blockdiagram window to ''wire-up'' the components ofour block diagram into a working instrument 3o that will be controlled by the front panel and display its output there as well. You can return to the block diagram window either by clicking on the exposed edge of that window or by choosing OPEN DIAGRAM from the Fl1 E menu. The first think you will notice is that your block diagram has changed. There are now four new objects in the block diagram window that were not there when we left. These block diagram representations of the front panel controls are called ''terminals''. These objects can be moved 35 about the block diagram just like the original block diagram elements. Notice that with the exception of the constants, each of these elements bears a resemblance to the control laid down on the front panel. The con- stants are distinguished by their numerical contents representing the values they hold. When a control is ''se- Iected on the front panel, its terminal equivalent on the block diagram is ''high-lighted'', and vice-versa. Before we begin wiring up the block diagram we need to move the control terminals into closer proximity to their points 4O of action. Using the grabber tool, click-drag each of the control terminals into positions similar to those shown in Fig. 54. We are now ready to begin to wire the elements into a working instrument. Note. If your original positioning of the block diagram objects differed significantly from our example, now is the time to rearrange your block diagram so that it more closely resembles Fig. 54. Eventually it should be relative easy to move about pieces of a wired diagram, but at present this is not possible. 45 To connect component in the block diagram you must use the wiring tool. This tool is represented in the tool palette in the upper left-hand corner of the block diagram window by the wire-roll glyph. The hot spot on this took is the end of the unrolled stretch of wire. Move the cursor to the tool palette and click to select the connector, when you move the cursor back into the active-window area your cursor will now be the connector tool. Connections are established between block diagram components by clicking in a hot spot of one element 5O and pulling the flickering line to a hot spot of component to which you want to connect. Figure 55 shows the connection between the slide-pointer control used to set number of iterations in the loop and the iteration-va- riable box in the loop (labelled N). The ''wire'' is represented by a flickering line during the connecting operation. Clicking in a hot spot of the terminal component will change the wire to a non-flickering representation of the connection. The interior ofobjects, orobject elements, represented by heavy black borders are ''hot'' anywhere 55 within their borders. Other objects, for example, the LINEAR FIT object, have localized hot spots representing specific inputs and outputs. Figure 56 shows the arrangement of hot spots for all available functions. Connecting wires can be ''forced'' to turn corners where the user wants them to for purposes of neatness and legibility ofthe diagram. Asingle clickwhile dragging a wire outside ofa structure or otherobject, will allow _l
gocr ep0242131-018
EP O 242 _3_ B_ the user to change direction at that point by 9O'. Pulling the wire again will allow the user to continue toward their target connection. Try using this technique to connect the pointer control to the iteration variable box on the lower loop structure. Often it is necessary to erase and redraw lines. To do this, use the grabber tool to select the line segment 5 to be removed. Remember where the grabber hot spot is located (at the junction of the thumb and first finger) when trying to select wires. With the line segment selected, pull down the EDIT menu and choose CLEAR. This action will erase only the selected line segment. Successive pieces ofthe line must be removed individually repetitions of this sequence of actions. Tiny pieces of line segment may be difficult to select. Sometimes line segments successfully selected and cleared appearto remain on the blockdiagram. Trial-and-error is the name 1o of the game here. Drawing replacement lines is identical to rawing new lines. Figure 57 shows the Fibonacci virtual instrument completely wired. Notice that the representations of the wires vary when they exit from various structures and functions. These changes reflect the data type being represented. Figure 58 shows the data types used by the pre- ferred embodiment and their representations in the block diagram. Notice that the data type output from the 15 LINEAR CURVE function in Figure 57 is a structure. This structure consists of two arrays representing the two functions to be graphed. One function is the connected data points, the second the linear fit calculated by the LINEAR CURVE function. You are now ready to return to the control panel and operate your instrument. Return to the front panel by finding the edge of the hidden window and clicking to bring it to the front. The front panel window has not 2o changed and should still appear as it did in Fig. 53. To operate the front panel choose the operator tool from the tool palette. The operator tool is the left-most glyph in the palette, with the appearance of a finger about to press a button. This tool's hot spot is at the end of the pointed finger. Use this tool to move the pointer on the pointer control to some value between 1 and 2O. Bring the pointed finger into proximity with the pointer on the control, click and drag the pointer to a new location. This action changes the value of the number of iter- 25 ations, N, to be performed by the upper control loop in the block diagram to which it is connected. To the right of the tool palette, above the front panel window, but below the menu bar are several controls (see Fig. 23). Only GO is currently operable, but this is all you need to set your instrument into motion. Use the operator tool to select the button and watch the indicators and graph. The top indicator represents the current value of the Fibonacci number generated in the top loop and the bottom indicator displays the square of the index used to 3o increment the loop. At the end of the loop a graph is displayed that contains two functions, the dark line is the connected data points representing the relation between a Fibonacci number and the square of its index in the sequence. The grey line represents the linear fit performed by the LINEAR CURVE function. You now have a working instrument, the construction of which has used many, but not all of the working features of the current prototype of the preferred embodiment. 35 To help you in your further exploration of the system, two sections follow which describe the use of the tools in more detail, and provide some hints forusing the sequence and ''for'' loop structures. These appendices and Fig. 56 are probably useful as ''cheat sheets'' forthe new userafterthe firstwalk-through with the Fibonacci virtual instrument. 4O (PICTURE)First Additional Explanation Section OPERATE TOOL. Use this tool to operate the controls on the front panel orset constants on the block diagram. (PICTURE)_ et or change control values - DOUBLE-CLICK in control to select region for typing in value 45 _ Begin execution CLICK in log or free box _ Set or change constants on block diagram - DOUBLE-CLICK in control to select region for typing in value _ Duplicate control, indicator, constants or structures 5O - SeleCt objeCt, then OPTlON-CLICK and DRAG Copy _ Clear object from front panel or Block Diagram - select object, then chose clear from edit menu GRABBER. Use this tool to select regions and move or resize objects. (PICTURE)_ Select control, indicator, or non-structure object 55 - CLICK anywhere on control or indicatoron front panel, or non-structure object on block diagram _Select structure - CLICK on bottom or right edge of structure _ Move object _8
gocr ep0242131-019
EP O 242 _3_ B_ - once selected, CLICK anywhere in selected region and drag _ Resize object - CLICK-AND-DRAG on lower right corner to desired shape and size (sometimes it is easier to see ''corne_' you are looking for if you select the object first) 5 WIRING TOOL. Use this tool to connect objects in the block diagram and to associate controls and indicators w(PICTURE)ith erminals on the front panel. _ Connect objects on the block diagram - CLICK on first object, draw wire to second object and CLICK - CLICK to make 9O' turn in wire 1o _ Choose Term Pane Pattern - drop TermPane (Format menu) - OPTlON-COMMAND-CLICK to go to next pattern _ Associate control or indicator and terminal - with chosen pattern visible in Term Pane, CLICK on control then CLICK on terminal 15 _ Check association of control and terminal - CLICK on control (associated terminal will be hilighted) _ Disassociate control or indicator and terminal - OPTlON-CLICK on control to be disassociated ALL TOOLS 2o (PICTURE)_ elect new tool - CLICK on tool _ Select menu item - CLICK on menu bar and drag down menu _ Drop Control, function or structure 25 - within dialog boxes, anytool may be used _ Move windows - CLICK-AND-DRAG on title bar ofwindow _ Title windows - DOUBLE-CLICK on title bar of one of the windows 3o _ Resize window to full screen - DOUBLE-CLICK on title bar of window _ Scroll windows - CLICK on arrows, or inside scroll bars _ Resize windows 35 - CLICK-AND-DRAG on grow-box in lower right corner ofwindow (PICTURE)Second Additional Explanation Section SEQUENCE STRUCTURE. Express explicit time dependencies, i.e., execute block diagram in frame O, then 4O f(PICTURE)rame 1 , hen frame 2. For best results, follow this sequence. 1) Determine how may frames you want - the first frame, O, is given when you drop the structure - to make a new frame COMMAND-CLICK with the grabber in the upper right corner and the frame 45 number will increment - do this for each frame desired - before filling any frames 2) Build block diagram in each frame - select frame to work on by CLICKING with the grabber in the upper left corner 3) Awire coming into a sequence can be used in any of the frames in the sequence by wiring to the dark 5O rectangle at the border. 4) Wires leaving a sequence can only come from one source. ''FOR'' LOOP STRUCTURE. Execute block diagram within loop N times. (PICTURE)So e hints. 1) To force the number of iterations connect a numeric control to the box in the upper left corner of the 55 structure containing an N, otherwise the loop will iterate ARRAYSIZE times 2) If new elementfrom an array is used at each iteration, the array must be outside ofthe loop. The number of iterations will be set by the number of elements in the array (unless overridden by setting N ?). Notice the array wire outside of the loop changes to a scalar wire inside the loop. _9
gocr ep0242131-020
EP O 242 _3_ B_ 3) The index (boxed i) in the structure is for your use, but need not be wired up 4) When building an array by producing a value at each iteration, the array is an indicator outside of the Ioop. Notice the scalar wire inside the loop, and array wire outside the loop structure. 5) To view intermediate values at run-time, place a scalar indicator inside the loop and connect it to the 5 wire carrying the value of interest before the wire leaves the structure. CASE (CONDITlONAL) SELECTlON STRUCTURE. Executes the block diagram chosen by the selector. (PICTURE)1) Same s sequence 2) Same as sequence 3) Awire coming in (same as sequence) 1o 4) All wires leaving a sequence must have sources (be wired) in every frame of the sequence. 5) The little box on the left-hand border is the selector box. At present, only numbers can be wired to it. ''WHILE'' LOOP STRUCTURE. Execute diagram within loop until test becomes false. (PICTURE)Som hints. 1) The little box with the Q in it should receive a boolean. When the input boolean is FALSE the loop will 15 terminate. If it is not wired the loop will be infinite! 2) The diagram in the ''while'' loop will always execute at least once. 3) The index (boxed i) in the structure is for the user's use, but it need not be wired up. 4) Arrays are not presently indexed when entering a while loop. 5) Same as ''for'' loop. 2O (PICTURE)Creating and Using Hierarchical Instruments The capability ofthe preferred embodiments to use previously implemented virtual instruments within new instruments is a fundamental and powerful feature of the system. This capability allows the user to construct 25 a hierarchy ofvirtual instruments by including previously implemented virtual instruments as ''parts'' in a new instrument. Tofacilitate the construction ofsuch complex instruments the usermust ''prepare'' the instruments to be included as components in an instrument hierarchy. This preparation includes defning the connections that will exist between the ''controlling'' (superior?) and ''Component'' (inferior?) virtual instruments and de- signing an icon to represent the ''component'' virtual instrument in the block diagram ofthe ''controlling'' virtual 3o instrument. Three items in the FORMAT menu support these features. ICON PANE, ICON EDIT PANE and CONNECTOR PANE. Choosing the ICON EDIT PANE option brings up the dialog box shown in Figure 59. Clicking on the ICON button brings forth two boxes, one large and one small (see Fig. 6O). The large box is the icon editing pane. Placing the cursor inside this box and clicking the mouse button produces a signal black square. Each square 35 will appear as a single black dot on the screen in the finished icon. Click dragging inside the box produces a line of these squares (see Fig. 61). Clicking on existing squares removes them. Operations inside this window are similar to the operation of MacPaint's FatBits mode. Actions performed in the large square are mimicked in the small square. The image in the small box is the default size ofthe object that will represent the instrument in the block diagram ofthe superior instrument. The dialog window has sev- 4O eral additional buttons formanipulating the editing pane. The CLEAR button clears the contents ofthe window. The INVERT button swaps black squares for white squares (and vice-versa) across the whole editing pane. REVERT replaces the current contents of the editing window with the last saved version of the icon. SAVE saves the current contents of the editing pane. Possibly useful if you have made a number of changes in the image and are uncertain, about the remaining changes you are about to make! The save version is the version 45 returned upon choosing the REVERT button. QUIT returns the user to the front panel window. (PICTURE)Displaying the Instrument Icon Chosing the ICON PANE option from the FORMAT menu puts the icon, Demo here, associated with the 5O current instrument in the upper left-hand corner of the front panel window (see Fig. 62). Choosing the same option again (now ''checked'' in the menu) removes the icon from the active window. (PICTURE)Using the Connector Pane to Establish Terminals 55 Creating instruments to serve as components within other virtual instruments requires that input-output connections or''terminals'' be established to provide a means fortransferring information between instruments. These connections are established through the mechanism of the CONNECTOR PANE. In the front panel window, selecting the CONNECTOR PANE option from the FORMAT menu places a 2O
gocr ep0242131-021
EP O 242 _3_ B_ small rectangular pane in the upper left-hand corner of the active window (see Fig. 63). Within this pane are several sectors, or ''terminals''. The configuration of terminals displayed in this pane represent the pattern of ''hot spots'' or terminal connections thru which information can flow between this instrument and other instru- ments, or instrument components. There are currently eight different patterns ofconnector terminals available 5 (see Fig. 64). The user can view the selection of available terminal patterns, when the wiring tool is in use, by holding down both the OPTlON and COMMAND keys, clicking the mouse button. Note. the cursor does not have to be in the connector pane to shift patterns. Each click will bring up a new terminal pattern. Once the eight available patterns are displayed, continuing to click will cycle through the selection again. The pattern displayed in the connector pane is the one available for ''wiring''. To make an association between a terminal 1o in the connector pane and a control (or display) select the terminal by clicking in it with the wiring tool. The selected terminal will turn black (see Fig. 65a). Move the cursorto the control to be associated with this terminal and click. The selected control will be encircled by the familiar moving dashed rectangle to indicate its selected status (see Fig. 65b). Moving the cursor off the selected control and clicking will de-select both the terminal and the control. The connected but de-selected terminal will now appear grey to show that it is connected (see 15 Fig. 65c). Connections can be established in the opposite order also. select the control and then select the terminal. At any later time the established associations between terminals in the connector pane and controls or displays on the Front panel can be examined by clicking on either a terminal or a control. The associated elements will both appear selected. Associations between terminals and controls can be disconnected by selecting the control, holding down 2o the COMMAND key and clicking. All connections can be severed by holding down both the OPTlON and COM- MAND keys and clicking (cursor must be the wiring tool). This action brings up the dialog box shown in Fig. 66. Choose OK and all associations will be removed and a new terminal pattern can be selected. Associations between terminals and controls can be disconnected by selecting the control, holding down the COMMAND key and clicking. All connections can be servered by holding down both the OPTlON and COM- 25 MAND keys and clicking (cursor must be the wiring tool). This action rings up the dialog box shown in Fig. 66. Choose OK and all associations will be removed and a new terminal pattern can be selected. Once you have created an icon and implemented terminal connections in the connector pane you can use this instrumentas a component in another instrument. From the blockdiagram window, pulling down the FUNC- TlONS menu and selecting the INSTRUMENTS option will bring up the familiar dialog box offering a choice 3o offile names in a window. Select the name ofthe instrument you wish to use in your block diagram. The dialog box will close and the icon associated with your chose virtual instrument will appear in the block diagram win- dow. At this point the virtual instrument object can be manipulated in exactly the same fashion as the functions you have used previously. In the next section a simple hierarchical instrument will be constructed in which two low-level instruments are operated through a controlling virtual instrument. 35 A(PICTURE)Hierarchial Virtual Instrument The simple hierarchical virtual instrument we will walk through calculates the total resistance for a parallel circuit with two resistive elements; 4O 1JR_ _ 1JR_ + 1JR_ SinCe We mUSt peffOfm SeVefal inVefSiOn OpefatiOnS tO Obtain R_, a fifSt inStfUment tO COnStfUCt iS a Simple inverter. Note. 45 1 . In building this hierarchial instrument it is assumed that the reader is familiar with the system as descri- bed in earlier sections of this patent application, and has successfully implemented the Fibonacci number generator. 2. To illustrate the construction of this instrument ''tiled'' windows are used to conserve time and space in generating the figures (see Fig. 26 re TILE option). 5o The first task, then, is to build a simple inverter. Figure 67 shows the front panel and block diagram for this instrument. In the block diagram, a constant and a variable quantity are the input to a division operator. The output from the division operation is fed to an ''indicator-only'' control. The constant has been set to one (in the block diagram). The slide control (with digital display) in the front panel determines the value to be in- verted. When these elements have been wired together, their correct function should be verified by setting 55 the slide control in the front panel window should display the inverse ofthe value specified on the slide control (see Fig. 68). You are now ready to establish the terminal that will accept a value from a higher-level instrument to set the slide control's value in this instrument, and to return the calculated value displayed here to the higher- Ievel virtual instrument. 2_
gocr ep0242131-022
EP O 242 _3_ B_ To obtain the connection pane in the front panel window, pull down the FORMAT menu and select CON- NECTOR PANE. This action will display a connector pane in the upper left corner of the front panel window. With the wiring tool selected, ''toggle'' through the terminal patterns by clicking the mouse button while holding down both the OPTlON and COMMAND keys until the two-terminal pattern is displayed (the left-most pane 5 in Fig. 63). Use the wiring tool to select the left terminal and then select the slide control. The terminal will blacken and the slide control will be surrounded by the familiar moving dashed line when they are selected (see Fig. 65) for the consequences ofa similar sequence ofactions). Click outside the control to complete the connection. The selected terminal will turn grey and the moving dashed line will disappear. The grey terminal is now ''connected'' to the slide control. The indicated portion ofthe square connector pane corresponds to the 1o input ''hot spot'' region on the instrument's to-be-built icon. Select the remaining terminal and connect it to the indicator-only control. In the same fashion as described above, this newly selected region will correspond to the output region on the instrument's icon. When both connections are complete, both terminals in the con- nector pane will be grey. At this time your screen display should appear similar to Fig. 64b. The next task is to create an icon to use with the inverter instrument. 15 Review the section on icon editing, set forth above, and create an icon similar to the one shown in Fig. 69. When your icon has been created, and saved, QUIT the icon editor. The inverter instrument is now ready to be saved. From the FILE menu select CLOSE and save the instrument with a descriptive name, like IN- VERTER. In yOUf neXt inStfUment yOU Will USe the INVERTER inStfUment tO CaICUlate 1lR_ f_ftWO feSiStanCeS in paf- 2o allel. Figure 7O shows the front panel and block diagram for an instrument to perform this calculation. At this point, only one new piece of information is necessary to construct this instrument. How to drop the new in- strument into the block diagram window. Pull down the FILE menu and select NEW. Pull down the FILE menu again and select OPEN DIAGRAM. Pull down the FORMAT menu and select TILE. From within the block dia- gram window, pull down the FUNCTlONS MENU and select INSTRUMENTS. A dialog box similar to the one 25 shown in Fig. 71 will appear. Unless you have done this before, or are using someone else's system, only one instrument will be avail- able. the INVERTER instrument you just created. Select this instrument by double clicking on the instrument name (or single clicking on the name and then single clicking on the OPEN button). The dialog box will disap- pear and your icon for the inverter instrument will appear in the block diagram window (see Fig. 72). This icon 3o now represents an object that can be moved about in the block diagram windowjust like the pre-defined func- tions and structures you have previously used. The ''hot spot'' for wiring to and from the instrument are the ones you specified when you connected the terminals in the connectorpane. As defined above in the connector panel, the left half of this icon's area accepts inputs to the control which it is wired and the right half provides output from the lower-level instrument's indicator-only control. 35 Verify that the virtual instrument correctly (see Fig. 7O) displays the sum ofthe inverses ofthe two resis- tances set with the two front-panel controls on the present instrument. Now use the icon editor to construct an icon for this instrument. Figure 51 shows the one used here. Finally, assign terminals to the controls and display in the connector pane. Use the 2-to-1 terminal pattern (second from the left in Fig. 64). Close and name the instrument. 4O Armed with the two low-level instrument you have created you are ready to implement a solution to the pfOblem Set. COmpUte R_ fOf tWO feSiStanCeS Wifed in pafallel. FigUfe 74 ShOWS the eXample SOlUtiOn tO the problem. Two controls are provided to provide the two resistances to be summed and an indicator-only control is implemented to display the answer. The block diagram window shows the two quantities specified by the front panel controls as input to the invert-and-add instrument (here named ADDlOVERR, see Fig. 75). The 45 OUtpUt ffOm thiS inStfUment iS inpUt tO the inVeftef tO yield the SOlUtiOn, R_, ShOWn in the diSplay Of the ffOnt panel window. Use several different values for input to verify the correct operation of your instrument (see Fig. 75). At this point you have created a hierarchical virtual instrument using native preferred embodiment func- tions and two user-created instruments. Hierarchical virtual instruments work. Libraries of instruments can 5O be created and selected instruments can be used as components in new instruments. With the principles dem- onstrated in this exercise you should be able to carry out your assignment and create libraries of instruments that perform functions useful to you as an end-user. Conclusion 55 (PICTURE) The method of the present invention, therefore, permits the computer-aided modelling of a process using graphical techniques which generally are more easily comprehended, especially by penons who do not pos- sess specialized skills in computer programming techniques. The use ofa computer-generated image ofafront 22
gocr ep0242131-023
EP O 242