Rapid user interface development with the script language Gist

Size: px

Start display at page:

Download "Rapid user interface development with the script language Gist"

Marilyn Byrd
9 years ago
Views:

1 RAPID USER INTERFACE DEVELOPMENT WITH THE SCRIPT LANGUAGE GIST G

2 Groningen Dissertations in Linguistics 8 ISSN 0928{0030

3 Rksuniversiteit Groningen Rapid user interface development with the script language Gist proefschrift ter verkrging van het doctoraat in de Letteren aan de Rksuniversiteit Groningen op gezag van de Rector Magnicus Dr S. K. Kuipers in het openbaar te verdedigen op donderdag 24 juni 1993 des namiddags te 1.15 uur precies door Gsbert Bos geboren op 10 november 1963 te 's-gravenhage. A C E G B D F H

4 1 e Promotor: prof. dr F. Zwarts 2 e Promotor: prof. dr J. van den Bos

5 Preface As more and more people are using computers routinely, not out of love for the machine, but because they need to, it becomes increasingly important to have easy-to-use, \intuitive" software. Consider, for example, the catalogue of a public library. Most people can nd a book if the catalogue is on index cards, but having all titles on-line in a computer has some advantages. For those who need it, it allows looking up books in ways not possible before { and if it is possible, sooner or later someone will need it. However, it forces every visitor of the library to use the computer instead of the cardtray. It's not only card-trays that are being replaced by computers, humans are ousted by them, too. Expert systems are deliberate attempts at making use of the knowledge of a human expert without the need of having him (her 1 ) around in person. Extracting the knowledge from a human expert and putting it in a form suitable for processing by a computer is only part of the problem, however. A computer program only makes sense if the people who would benet from it are able to use it. Asking someone a question is something we are all very experienced in, but asking a computer a question is a completely dierent thing. This doesn't imply that the task of the designer of a \user interface" is to make the computer mimic a human, far from that! Occasionally talking to a computer is indeed the best way of interacting, but more often other ways are much more ecient. The computer can display pictures, charts and other things and the designer of the interface would do well to make use of these capabilities. That the task of creating a good interface is at least as hard as that of making the computer solve a particular problem is 1 For ease of reading, only \he" has been used in this text. \He" should be taken to mean \he or she" where appropriate. v

For those who need it, it allows looking up books in ways not possible before { and if it is possible, sooner or later someone will need it.

6 Preface 2 Another reason for rewriting software could be that the existing software has grown so inecient from many patches that a complete overhaul saves more than it costs. An example is the Oberon System, an operating system, window system and programming language that together provide not much more than other systems, but in a much smaller, more consistent and faster package. Of course, no such argument can apply to user interface development systems. proven by the fact that on average about half of the time of writing a program is spent on the interface part (assuming it uses graphics of some kind). Since the interfaces of dierent programs often have a lot in common, this is an area where specialised tools can make life a lot easier. A number of such tools have been introduced already, most of them meant to be used in conjunction with a particular \window system", such as Microsoft Windows on the PC, or X Windows on unix machines. Tools can be targeted at several types of developers. Most often they are meant to be used by professional programmers; these tools usually give access to everything the window system has to oer. Some tools are made for people who only occasionally want to program. These tools oer only limited access to the window system, but they are much easier to use. One of the most important lessons that every programmer has to learn, is \not to reinvent the wheel". This means that using pieces of software that are available is often better than writing them oneself. This lesson can be extended to the use of tools. Use available tools, preferrably the simplest, since it saves you so much time. Only when you are sure you need more than a particular tool can provide should you switch to another or to doing it yourself. 2 In spite of the sound advice in the last paragraph, part of this book describes my eort to do just that: reinvent the wheel. Of course, there is a reason and there is a result. The reason is, that to fully understand user interface technology it is best to treat it as children do an alarm clock: take it apart and put it back together again. The result is a perfectly usable tool, with its own advantages and disadvantages. Although this book talks about human factors and the design of interfaces, it is not about designing interfaces per se, only about implementing them. Designing involves choosing metaphors, creating pictures, choosing colours, selecting keystrokes and making a lay-out for the screen. It is very dicult to do it right and there is no guarantee that you will be able to create nice looking and usable interfaces after reading this book. On the other hand, the tools described in this book are designed to make the task of creating and improving interfaces easy, so you won't lose much time if you have to try a few times before it comes out right. Acknowledgments During the time that I did the research for this thesis and while I was writing it I learned a lot more than I was able to put vi

package. Of course, no such argument can apply to user interface development systems.

7 Acknowledgments into the text. I want to thank my colleagues and ex-colleagues of the section Alfa-informatica, Jan de Vuyst, Harry Gaylord, George Welling, Peter Blok, Yvonne Vogelenzang, Erik Tjon, Erik Kleyn and Gosse Bouma, for many inspiring talks. I want to thank Gosse especially, for his critical reading of a draft version and for the shower of useful articles and books. Erik Tjon was the rst user of Gist, his enthusiastic eorts inspired many small improvements and he also made the rst non-trivial interface. I learned a lot from Kees de Vey Mestdagh, especially about writing articles, going to conferences and the like. I am indebted to the members of Phonk for keeping my curiosity alert and oering many other interesting topics. Groningen, 23 November 1993 vii

Bouma, for many inspiring talks. I want to thank Gosse especially, for his critical reading of a draft version and for the shower of useful articles and books.

8 Groningen Dissertations in Linguistics (Grodil) 1. Henri ette de Swart (1991). Adverbs of quantication: a generalized quantier approach. 2. Eric Hoekstra (1991). Licensing conditions on phrase structure. 3. Dicky Gilbers (1992). Phonological networks: a theory of segment representation. 4. Helen de Hoop (1992). Case conguration and noun interpretation. 5. Gosse Bouma (1993). Nonmonotonicity and categorial unication grammar. 6. Peter I. Blok (1993). The interpretation of focus: an epistemic approach to pragmatics. 7. Roelien Bastiaanse (1993). Studies in aphasia. 8. Bert Bos (1993). Rapid user interface development with the script language Gist. To appear in 1993: 9. Wim Kosmeer. Barriers and licensing. 10. Jan-Wouter Zwart. Dutch syntax: a minimalist approach. 11. Sietze Looyenga. Syntax and semantics of nominalizations. 12. Ale de Boer. VP-anaphora in contemporary English. 13. Petra Hendriks. Comparatives in categorial grammar. 14. Mark Kas. The semantics of verbs. colofon Opgemaakt met TEX3.14t en Postscript 2.0, Illustraties met Xg Lay-out: Bert Bos Lettertypes: Gladiator & Helvetica Printer: VariTyper 600 Oset-druk: Universiteitsdrukker, Groningen Oplage: november 1993

Gosse Bouma (1993). Nonmonotonicity and categorial unication grammar. 6. Peter I. Blok (1993). The interpretation of focus: an epistemic approach to pragmatics. 7. Roelien Bastiaanse (1993).

9 Contents Preface Contents v viii 1 Introduction Sociology Psychology Ergonomics Audience Preliminaries Denitions Workstations Interaction UIMS and UIDE Conversation analysis Resources Eciency Client-server model Hypertext/hypermedia Overview 12 2 Human factors & gui's The user's perspective Crammed displays The use of colour Interesting failures Classication of interfaces Goals of user interfaces Direct Manipulation Virtual Reality Windows 28 viii

7 Overview 12 2 Human factors & gui's 15 2.1 The user's perspective 16 2.1.1 Crammed displays 16 2.1.2 The use of colour 17 2.1.3 Interesting failures 18 2.

10 Contents 2.6 GUI's Interface elements Windows Boxes & menus Icons Buttons Images Sliders Prompts (text elds) & editors Lists File selectors Accelerators Resources Congurability, adaptability & intelligent interfaces Congurability Adaptability Conversation and role playing Small scale intelligence The process of designing UI's Style guides and guidelines 47 3 User interface development systems Software development techniques Specication vs. iterative design Prototyping Object-oriented programming Lazy & eager evaluation Non-deterministic design Some history Current systems Advantages & disadvantages Toolkits Interactive systems Interactive script-based systems Non-interactive script-based systems 63 4 Gist Separation of interface and application Building an interface The script language User input Messages among objects Physical and synthetic events Actions 78 ix

9.4 Small scale intelligence 44 2.10 The process of designing UI's 45 2.11 Style guides and guidelines 47 3 User interface development systems 49 3.1 Software development techniques 50 3.1.1 Specication vs.

11 Contents Expressions Calling external programs An example that uses cloning Modelling the application Setting global defaults Conguring & extending Gist Portability Advantages, disadvantages, possible enhancements Coupling of application and interface Script language On-line help & error handling Interactive design Possible enhancements 93 5 The implementation of Gist Flow of control Datastructures The \main" module The \parsele" module The \scan" module The \parse" module The \parseaux" module The \actions" module The \classes" and \classes.def" modules Conguration 106 A Gist syntax 111 A.1 Terminal symbols 112 A.2 Rules 113 B Gist example 119 B.1 Application model 120 B.2 Dialog boxes 121 B.3 Main window 122 B.4 Flagging impossible moves 123 B.5 Using higher level objects 123 Bibliography 125 Samenvatting 131 Index 135 x

$8.4 Interactive design 93 4.8.5 Possible enhancements 93 5 The implementation of Gist 95 5.1 Flow of control 95 5.2 Datastructures 97 5.3 The \main" module 100 5.4 The \parsele" module 101 5.$

Why participation works

Why participation works Full title Why participation works: the role of employee involvement in the implementation of the customer relationship management type of organizational change. Key words Participation,