« PreviousContinue »
2. The number of data sources and points of distribution to be encompassed by the network and their locations.
3. The volume of information (in terms of messages and lengths of messages) which must flow among the various locations.
4. How soon the information must arrive to be useful. What intervals the information is to be transmitted and when. How much delay is permissible and the penalty for delays. 5. The reliability requirements with respect to the accuracy of transmitted data, or system failure and the penalty for failure.
6. How the total system is going to grow and the rate of growth.”
(Probst, 1968, p. 19).
More specifically, overall system design considera
tions with respect to information flow requirements typically involve calculations of average daily volume of message and data traffic, peak loads anticipated, average message length, the number of messages to be transmitted in given time intervals, total transmission time requirements, and questions of variable duty cycles for different system and network components.3.23
Examples of relatively recent developments in this area include RADA (Random Access Discrete Address) techniques 3.24 and a "hot-potato" routing scheme for distributed networks.3.25 A continuing R & D challenge in terms of scientific and technical information services has been posed by Tell (1966) by analogy with the techniques of input-output economics.3.26 Of major concern is the problem of high costs of communication facilities necessary to meet network information flow requirements.3.27
4. Input-Output, Terminal Design, and Character Sets
The area of input-output, especially for twodimensional and even three-dimensional information processing, is currently receiving important emphasis in overall information processing system design. One reason for this, as we have seen, is the increased attention being given to remotely accessed, time-sharing, or man-machine interaction systems. In particular, as noted by Tukey and Wilk: "The issues and problems of graphical presentation in data analysis need and deserve attention from many different angles, ranging from profound psychological questions to narrow technological ones. These challenges will be deepened by the evolution of facilities for graphical real-time interaction." (Tukey and Wilk, 1966, p. 705).
4.1. General Input/Output Considerations
Since a multiplicity of input and output lines are assumed for a variety of types of information to be processed (including feedback information from users and from the system itself), development requirements with respect to both equipment and sortware processing operations include batching of various input units, buffering of at least some types of input (as required, for example, to provide necessary reformatting), and multiplexing of input operations. Such considerations also apply even more forcefully to interfaces between the various nodes of a network involving more than one type of participating system.4.
Format control is typically needed both into and out of the system, preferably under dynamic program control. The format control subsystem, by means of address storage registers or other techniques, should enable the input data itself to determine where it should go in storage, and other means of "self-addressing" should be provided
without the need for elaborate or inefficient programming and related software requirements.4.3
The overall output capability design should provide ability to reformat conveniently and efficiently 4.4 as well as to select certain character sequences. Because of the variety of equipments needed for various tasks, provision should be made for reversal of the bit order of input and output data so that either high or low order bits can be processed first. In the case of displays, special provisions may be required to prevent overlapping of symbols.4.5
Related to format control is the question of variable byte size for input and output. For the future, system design will require ASCII (American Standard Code for Information Interchange) code sorting and ordering capabilities, but in many circumstances it will also be necessary to handle collapsed subsets of ASCII and other codes, longer byte lengths such as 10- and 14-bit codes for typesetting, and even longer codes for monotype, numeric process control, data logging, and equipment control.
Analog-digital and digital-analog convertibility is needed for experimental applications in source data automation, measurements automation, map analysis, map and contour plotting, pattern processing, and the like. One example of convergent efforts in the field is provided by Ramsey and Strauss (1966) who discuss interrupt handling in the area of hybrid analog-digital computers as representative of more general on-line scheduling problems. For some of these investigations, at least virtual real-time clocks will be needed.4.6 This implies processor main frame and transfer trunks versatile enough to handle these requirements whether implemented by software or built into the hardware.
Another important requirement is for versatile and varied graphic input and output capability,
including light pen, microfilm, FOSDIC-type scanning, mark-sensing, OCR (Optical Character Recognition), MICR (Magnetic Ink Character Recognition), color-code input (such as Lovibond color network), and three-dimensional probe data in (see the first report in this series), and largevocabulary character and symbol generation; diagram retrieval, construction and reconstruction, and perspective or three-dimensional projection capabilities out (as discussed in the second report in this series). Photographic and TV-type input and output with good resolution, hard-copy reproduction capability, varying gray-scale facility, and at least the possibility of handling color input or output display techniques will be required in future system design.4.7 Audio input-output capabilities should include dataphone, acoustic signal inputs, and voice, with speech compression on output, requiring controlled timing of "bursts" or "slices."
In many system design situations, we should be able to switch peripheral equipment configurations around for special purposes and we may need to have multiple access to various types of peripheral devices simultaneously during the same processing run, e.g., to be able to shift between character recognition and graphic scanning tasks for input of material where text and graphics are intermixed.
Related to these problems of input, output, and on-line responsiveness (especially for clients involved in problem-solving applications), is the concept of graphical communication generally.4.8 This presupposes, first, a suitable language for the exchange of both pictorial data and control information between the designer and the machine, and secondly, provisions for the dynamic manipulation of data and controls.4.9
Recent programming techniques under investigation for the display of two-dimensional structure information are exemplified in work by Forgie,4.10 by Hagan et al. (1968) 4.11, and at the University of Michigan (Sibley et al., 1968).4.12 Then there is the DIALOG programming system developed at the ITT Research Institute in Chicago for graphical, textual, and numeric data input and display, online and offline programming facilities, and hard-copy options. (Cameron et al., 1967). A special feature is a character-by character man-machine interaction mode, so that the programmer may use only those input symbols that are syntactically correct. For more efficient machine use in production-type operations, a DIALOG compiler for the IBM 7094 has been prepared following the "Transmographer" of McClure.4.13 (McClure, 1965).
Then we note that "in the area of displays, determining the information to be displayed and generating the procedures for retrieval and formatting of the information are the difficult problems." (Kroger, 1965, p. 269). Further, as of today, "too many systems are designed to display all the
data, and not to display only the data needed for the decisions the system is called upon to make." (Fubini, 1965, p. 2).
In general, the client of the on-line, graphical input-output, and problem-solving system needs convenient means for the input of his initial data, effective control of machine processing operations, effectively instantaneous system response, displays of results that are both responsive to his needs and also geared to his convenience, and handy means for the permanent recording of the decisions and design choices he has made.
With respect to these client desiderata, the identifiable R & D requirements relate to keyboard function key overlay design; 4.14 improvements in both problem-oriented and client-oriented languages for man-machine communication and interaction;4.15 fast, high-resolution, flicker-free display generation; 4.16 ability to selectively emphasize various areas of display,4.17 further development of the combination of static displays (such as maps) with computer-controlled dynamic displays,4.18 and rapid responsivity of the system to feedback from the client.
Since remote, reactive terminals are an increasingly important factor in systems involving dynamic man-machine interaction, the question of design of remote inquiry stations and consoles necessarily raises problems of human engineering for whose solution there is inadequate experimental cost-benefit, and motivational data 4.19 available to date. Also involved are questions of acceptance and interactive response by the client to feedback outputs from the system, including requests for further information or additional inputs and display of re-processed results.
4.2. Keyboards and Remote Terminal Design
Where graphic input and output facilities are to be available to on-line users, there are unresolved questions of interrelated and interlocking system and human factors. How clumsy are light pens or pointers to use? Are they heavy or difficult to aim? 4.20 Should light-pen imputs be displayed a little to the left or to the right of the actual lightpen location so that the active part of the input is not blocked from view by the moving light-pen itself? 4.21 Can flicker-rate be kept to a tolerable level without undue and costly regeneration demands on a multiply-accessed central processor used by the many clients, or must the remote terminal have storage and display re-generation capabilities at added cost and design complexity? 4.22 For graphic input and display should the input surface be flat, upright, or slanted? 4.23
It has been pointed out, in the case of the recent development of a solid state keyboard, that "the requirements of today's keyboards are becoming more complex. Increased reliability and more flexibility to meet specialized demands are essential. Remote terminals are quite often operated by
relatively untrained personnel and the keyboard must be capable of error-free operation for these people. At the same time it should be capable of high thru-put for the trained operator as will be used on a key tape machine.
"Some of the limitations of existing keyboards are: ● Mechanical interlocks which reduce operator speed.
Excessive service (increasingly important for remote terminals).
Contact bounce and wear of mechanical switches.
● Non-flexible format." (Vorthmann and Maupin, 1969, p. 149).
For automatic typographic composition applications, it is emphasized that "the application of computers to typesetting only emphasizes the scope and the need for a radical re-thinking on keyset design," and that, although "one may imagine that the keyboard is a relatively simple piece of equipment . . . in fact, it presents a unique combination of mechanical, electrical and human problems." (Boyd, 1965, p. 152). Current R & D concerns with respect to keyboard redesign involve consideration of principles of motion study as applied to key positioning, key shape, key pressures required, and the like.4.24
Nevertheless, it is to be emphasized that "input-output devices are still largely the result of an ingenious engineering development and a somewhat casual and often belated attention to operator, system attachment, and programming problems' and that ". . . no input-output device, including all terminals combined, has yet received the careful and competent human factors study afforded the cockpit of a military aircraft." 4.25 (Brooks, 1965, p. 89).
Beyond this are questions of design requirements for dynamic on-line display. Thus we are concerned with requirements for improved remote input console and terminal design.4.26 Relatively recent input-output terminal developments, especially for remote consoles or dynamic man-machine interaction, have been marked by improved potentialities for two- and even three-dimensional data processing and by further investigation of prospects for color, as discussed, for example, by Rosa (1965), 4.27 Mahan (1968) 4.28 and Arora et al. (1967), among others. Van Dam (1966) has provided an informative state-of-the-art review of such scanning and input/output techniques. Vlahos (1965) considers human factor elements in three-dimensional display. Ophir et al. (1969) discuss computergenerated stereographic displays, on-line.4.29
In the area of input-output engineering and system design, what is needed for more effective manmachine communication and interaction will include the provision for remote consoles that are truly convenient for client use. Hardware, software, and behavioral factors are variously interrelated in
The desirable design specifications for remote inquiry stations, consoles, and terminals and display devices as discussed in the literature variously include: economy, dependability, and small enough size for convenient personal use.4.31 Some misgivings continue to be expressed on this score. Thus, it is reported that Project Intrex will consider the design of much more satisfactory small consoles 4.32 and Wagner and Granholm warn that "at the moment, it is difficult to predict whether remote personal consoles can be economically justified to the same extent that technological advances will make them feasible." (1965, p. 288). Cost certainly appears to be a major factor in the limited nature of the use that has been made of remote terminals to date.4.33
A second requirement is for the provision of adequate buffering facilities including, for at least some recent systems, capabilities for local display maintenance.4.34 From the hardware standpoint, it is noted that "the major improvements in displays will be in cost and in the determination and implementation of the proper functions from the user standpoint. The cathode-ray tube will probably be dominant as the visual transducer for console displays through 1970, but there are several new techniques for flat-panel, digitally addressed displays presently under development that may eventually replace the CRT in many applications. The advances in memory and logic component technologolies will permit significant improvements in the logic and memory portions of console displays." (Hobbs, 1966, p. 37).
Other features that are desirable may include a capability for relatively persistent display, for example, up to several hours or several days,4 and the capability, as in the Grafacon 1010 (a commercially available version of the RAND Tablet) for the tracing of material such as maps or charts to be superimposed on the imput surface, or the Sylvania Data Tablet ET 1, which also allows a modest third axis capability.
As in the case of system outputs generally, hard-copy options are often desired through the terminal device. For example, the console "should have a local storage device on which the user can build up a file of the pieces of information he is retrieving, so that he can go back and forth in referring to it. It should have means of giving him low-cost hard copy of selected material he has been shown and temporarily stored." (King, 1965, p. 92).
The use of markers and identifiers for on-line text editing purposes should be simplified or eliminated to the maximum extent possible. If, for example, elaborate line and word sequence identifications must be used both by the machine system and by the client, then the virtues of machine processing for this type of application will be largely lost. Such systems should not only be easy to use, but easy to learn how to use.
An important question to be asked by the system designer is whether the output responses will be of the types and in the formats that the client will expect to receive. It is noted in particular that "the closer the correspondence of the computer output is to the methods of presenting textual and tabular material familiar to the user, the greater his information absorption rate will be." (Morenoff and McLean, 1967, p. 20). Thus, in engineering applications, for example, the client may want results to be displayed in a familiar format such as a Nyquist plot.4.36 Similarly, in operations on files or data banks, the user should be able to structure and sequence files and subfiles for display and selection to suit his own purposes.
For effective online operation, the system should provide responses within the reading rates of typical users, and with good resolution, little or no flicker, and with both upper and lower case. 4.38 A some what more specific and stringent list of remote terminal desiderata is provided by Licklider, with particular reference to the requirements of multiple access to the body of recorded knowledge. His desired features include, but are not limited to, color, or at least gradations of gray scale; 4.39 terse or abbreviated modes of expression to the machine with full or "debreviated" response; 4.40 selective erasibility of the display by either program or user command, 4.41 and capabilities such as the following: "Shortly thereafter, the system tells me: 'Response too extensive to fit on screen. Do you wish short version, multipage display, or typewriter-only display?"." (Licklider, 1965, p. 50). Another design criterion affected by the factor of client convenience is that of the extent of display on the console face of meta-information.4.42
Continuing needs for technological developments have also been indicated for improved terminal and output display design. Examples include the development of new, fast phosphors and other materials,4.43 use of analog predictive circuitry to improve tracking performance,4.44 and variable sequencing of processor operations in computation and display.4.45 A number of advanced techniques are also being applied to large-screen displays,4.46 although some continuing R & D difficulties are to noted.4.47 Multiplexing of graphic display devices may also be required.4.48
Returning, however, to the human behavioral factors in input-output and terminal design, we note that man-machine relationships require further investigation both from the standpoint of human engineering principles and also from that of attitudes of clients and users,4 4.49 that there are continuing requirements for research and development efforts on both sides of the interface 4.50 and that, in all probability, "industry will require more special prodding in the display-control area than in the other relevant areas of computer technology." (Licklider, 1965, p. 66).
We note further an area of R & D concern that will recur in many other aspects of information
processing system design and use, and especially in information selection and retrieval applications. More specifically: "The major problem today in the design of display systems is that we cannot specify in more than qualitative terms such critical criteria as 'context' and 'meaning'." (Muckler and Obermayer, 1965, p, 36). Swanson adds: "Other restrictions derive from the need of programs to solve hidden line problems, to recognize context, and to make abstractions." (Swanson, 1967, p. 39). Finally, we note the specific problem in documentary and library applications that large character repertoires are important to input and output representations of various levels of reference and emphasis in technical texts (especially, for example, in patent applications with multilevel referrals to accompanying drawings and diagrams) and to delineation of different types of possible access points in indexes and catalog cards.4.51 In addition, a wide variety of special symbols and/or exotic alphabets are typically employed in texts dealing primarily with mathematical, logical or chemical subjects. A text written principally in one particular language and alphabet may frequently use the alphabet set of one or more other languages, as in reference to proper names, citations, and quotations.
4.3. Character Set Requirements
For multiple-access systems "most creative users want large character sets with lower-case as well as capital letters, with Greek as well as Latin letters, with an abundance of logical and mathematical signs and symbols, and with all the common punctuation marks." (Licklider, 1965, p. 182). In addition, subscripts, superscripts and diacritical marks may be required.4.52
Attacks on problems of character-set requirements for output begin with on-line printer variations to provide larger output character-set vocabularies at the expense both of output printing speed and of prior input precedence-coding and/or of processor programming. Larger characterset vocabularies are provided both by photocomposition techniques and by electronic charactergeneration methods, but again at the expense of either pre-coding or programming requirements.4.53
It should be noted, of course, that the internal language of most general-purpose information processing systems is limited to no more than 64 discrete characters, symbols, and control codes. Thus there must be extensive provision for multiple table lookups and/or for decoding and reencoding of precedence codes or transformation sequences, on both input and output, if any internal manipulations are to be performed on the textual material. In general, the larger the character set, the more elaborate the precoding and/or programming efforts that will be required.
Then there is the problem of setting up keyboard character sets that are adequate for application requirements and yet within reasonable
human engineering limitations. One proposed solution, the use of keyboard overlays and control codes to enable rapid shifting from one character subset to another, is exemplified by developments at Bunker-Ramo.4.54 Another possible solution to the character set problem as related to human engineering factors that is receiving continuing R & D concern is to provide multiple inputs via a single keystroke, such as "chord" typewriters or Stenotype devices.4.55
Regardless of what may be available through various conversion, transliteration, or translation processes on either input or output, there remains the question of the effects of internal character set upon sorting, ordering, filing and interfiling operations for a specific processor-storage system. For example, "other control elements which are frequently required in the design of information systems are special sorting elements. In a directory of personal names, such as those which might be found in an author bibliography, if names beginning with 'Mc' and those beginning with 'Mac' are to sort together, then special sorting codes must be entered into the computer for this purpose. (Austin, 1966, pp. 243-244)4.56
The size of an adequate character set is a particularly critical problem in at least two significant areas: those of automatic typographic-quality typesetting and of library automation.4.57 Complex character set requirements are also to be noted in such multiple-access applications as computeraided-instruction (CAI).4.58 Avram et al., considering automation requirements at the Library of Congress, stress that "keyboard entry devices must be
designed to facilitate the input of a variety of languages and diacriticals." (1965, p. 4). These authors point out further that "if the problems associated with the design of input keyboards and photocomposition printing devices can be resolved for the multiplicity of alphabets, there still will remain the formidable task of searching in a machine file which contains them." (Avram et al., 1965, p. 89). Similarly, Haring (1968) points out that the 128 symbols provided in the ASCII code is inadequate for the augmented catalog under development at Project INTREX.4.59
The very number and diversity of varied but realistic cataloging, filing, and search considerations in terms of character-set and sort-order requirements that exist today may indeed surprise the typical computer scientist facing library automation implementation problems. Nevertheless, particular problems of sorting, filing, and reassembly orders in terms of practical usage needs and acceptability to the clients of mechanized systems and services should be subjects of concern to designers of machine languages, machine character-sets, and of the processors as such.4.60
The even more difficult case of extensive mathematical, chemical, and other special symbols desired on output imposes additional hardware requirements, whether for high speed printers, photocomposition devices, or character generators.4.61 This, then, is the area that has been called "Caligraphy by Computer." 4.62 A final example of unusual character set requirements is provided by "Type-A-Circuit" developments.4.63
5. Programming Problems and Languages and Processor Design Considerations
The questions of design and development of appropriate programming languages and of processor design are obviously pertinent to all of the operations shown in Figure 1. As of 1967-68, however, special emphasis in terms of research requirements lies in three principal areas: user-oriented input, response and display languages; symbol manipulation languages capable of handling arrays of multiply interrelated data, and increasing interpenetration of hardware and software considerations in both system design and system use.5.1 For example, in the operations of developing processing specifications from client requests for service (Boxes 8 and 9 of Fig. 1), we need: new and more powerful problem-oriented languages; versatile supervisory, executive, scheduling, and accounting programs; hierarchies of programming languages; multiprogramming systems; improved microprogramming; new approaches to increasing interdependence of programming and hardware; and more versatile and powerful simulation languages.
The overall system design requirements of the future indicate R & D concerns with programming languages, and especially with hierarchies of such languages at the present time.5.2 Controversies certainly exist as between advocates of more and more "universal" languages and proponents of problem-oriented or user-oriented machine communication techniques.
5.1. Programming Problems and Language
Continuing R & D requirements for programming language improvements represent two contradictory requirements: on the one hand, there is recognized need for increasingly universal, common-purpose languages compatible with a wide variety of systems, hardware configurations, and types of applications; and on the other hand for hierarchies of language systems. In addition, a number of special requirements for more flexible, versatile and powerful languages are just beginning to emerge, especially in such