- For the afternoon session "Interfacing: codex versus electronic text", at the conference "The Future History of the Book", Koninklijke Bibliotheek, the Hague, november 8th 2002 -
I found this text amongst the HTML-files on my own website. It was not linked from the index-page. I do recall being part of a conference at the KB – I also have a memory of listening to a lecture by Robert Darnton – but I can't remember that I talked about the things I mention in the text below.
Questions raised | intro
Since the rise of hypertext and the World Wide Web a lot of questions regarding the user interfaces for computerdata have been raised.
Some of those questions, I think, could very well have been based on a flawed understanding of how people interact with a computer, how they interact with information that comes to them from a computerscreen, and a flawed understanding of how people interact with information.
I will try to make this clear by putting up a few propositions and - maybe controversial - opinions. I will react directly to both the propositions brought forth by the organisers of this conference:
The interface of the book allows a wealth of variation within a familiar and predictable standard. Such an easy-to-use and versatile interface is missing in the digital realm.
I would suggest that either such a standard is already existant (the WIMP - Windows, Icons, Menu, Pointer - paradigma that led to the GUI - Graphical User Interface -, that inherited a.o. Windows); I would also suggest that the good thing about computerreadable information is exactly that it can be represented by different interfaces.
Unlike printed books, internet publications are like black boxes: there is almost invariably more content than you get to see on screen. The development of user interfaces that allow a transparent look at the nature and extent of the content similar to that of the book is particularly urgent.
I will suggest that this is not a real problem; that a book is also in a lot of ways a black box; and finally that - arguing from the point of view of the user - such a transparant look is not at all urgent.
intro for the booklovers
For the hardcore booklovers - let it be clear that I love books; and let it be clear that for sheer reading pleasure, a well-designed book is the best thing you can get.
This has to do with material aspects: there's no light shining directly into your eyes, the paper doesn't flicker at, say, 95Hz, most books are more portable. lighter and easier to handle than a laptop, and there's no distraction that comes directly from the interface of this reading machine. And - when the book is yours - you can scribble notes in the margin. (Why that is more useful than jotting down notes in simpletext or making annotations in hyperlinks is another question, not to be answered here).
On the other hand: when I write, or do research, or when I'm busy preparing classes, I prefer computerreadable information. The reasons for it are largely pragmatic. (Possibilities of text-search, copy-pasting, mail important bits to students, steal text for the website, quick research of context.).
As you all know: books are no longer the dominant medium for the transmission of information. Not in the western world; and possibly also not in a lot of poorer countries). This has a.o. economical reasons. But the subject under scrutiny here is the interface. Not the dynamics of media.
the second question
One problem that's been haunting the question of designing good interfaces for cd-roms, hypertext and websites has been that almost all of the interface give the user no idea or feeling of the amount of available information. This is supposedly a basic problem for information stored on a computer.
I would like to maintain that this is a pseudo-problem.
it's the time spent
Behind this problem lies the idea that when you hold a book you do have some idea of the amount of available information. But is this really the case? You get an idea of the amount of pages, when you open the book and quickly scan the pages, you see how many pages there are, how big the print is; if you put in some extra effort you might get an idea of the content by looking at the titles of chapters. Still going further you could read a few sentences and get an idea of the style and the density of the prose.
Compare this with how one interacts with a cd-rom or website. You take a quick look at the navigation, you see the graphical design, you click through a few pages without reading, retrace your steps, maybe you look at a sitemap, or some peripherals.
But then - do you have an idea about the amount of valuable information? The information value will unfold in time. There's no way of knowing it in advance. It's in the time of working with a single 'storage-medium' that one gets an idea of the value of the information.
What you would like to catch a glimpse of - through an interface - is the amount of information. Books nor cd-roms or websites do this. Or better, they do it, in some ways, thanks to culture-specific conventions, genre &c. It has nothing - or at least not much - to do with a general interface. What one can get to know quickly, from books, and what one can get to know from cd-roms, dvd's and websites as well, is the amount of available data. How many pages and sentences; how many megabytes.
Why would you want an interface that gives you an idea of the amount of available 'data'? (Couldn't that be easily measured in amount of words, or amount of megabytes?) What, in the end is important, is what one gets out of the information. It's a clichee: some big books are utterly superfluous, some short articles contain more insights that long books.
Information and knowledge do not stop at the edge of a storage medium (a book, a cd, a site, a video). This is increasingly the case in a world in which electronically linked information has become dominant, but it really has always been the case. What the experience of designing websites and computerinterfaces, and working with them, has made clear, is that - in designing - one should think from the perspective of how users interact with information (or, of course: how they should interact with it).
In an information world in which documents become more and more interlinked, who needs to know the 'size' of information inside some well-defined limits?
Only (traditionally told) stories are self-contained. (Games btw might be another genre where this is the case). All other genres of information transmission point to other places, other texts. Studying one book asks for discipline; asks for the discipline of not to take to hand that other book, not to search for further information. But in a more positive sense we could call this - hmm - following the path that leads you to the real information, and to reflexive knowledge. (but you have to make sure you plot your way, that's true).
But I don't want to go to far into this wildly fascinating territory (that concerns questions like: how do we learn? what's the best way to learn? how do we attain knowledge and insight?) For now I'd like to stop at the fact, that people interact with most books (excepting novels) in more or less exactly the same way as they interact with information on the web: they surf from page to page, search, read a bit, follow links, search, read a bit, and sometimes read a whole article. People are finding ways of navigating the information sea.
the good thing about computer-readable data
The good thing - yes, maybe one of the best things - of computer-readable data, is that we can look at those data in different ways. Literally by using different interfaces, and different programs.
Programmers, nerds, software - critics, activists and some academics are probably the only groups who sometimes switch between Command Line Interfaces and Graphical User Interfaces. (More people should do it, I think. To raise their understanding of what they're dealing with (both the computer and the information that gets to them). Some members of these groups will actually defend the idea that the CLI of unix is really, in the end, more user-friendly than all of the GUI's. We could dismiss this as the crackpot ideas of a minority. On the other hand, nothing is easier as serving data in such a way that they can be read in different ways. It is what computers are about. And the whole WWW is a good example. The problem lies a.o. with the people who have no understanding at all of what a browser is and what an application actually does. MacOSX is in that sense a little blessing.
This advantage has - in daily reality - almost disappeared. Most people mistake the horrible Windows-interface for the computer.
Of course, saying this is evading the question of what a good interface is. The answer might be: there are several good interfaces; for some a CLI is handier (programming for instance, or configuring a web server), for some tasks you'd like a GUI. And sometimes (sometimes?) you'd rather click through the folders of a cd-rom than deal with the niceties of the graphical design.
Ask yourself this question: what are the reasons for some people to prefer straight code-editing to WYSIWIG-editing? Why do others prefer a graphical presentation of programming (like in MAX MSP). Why the hell do some people still prefer Emacs, or even VI? When is a GUI an empowerment of the user, when is it a pure nuisance to the user?
"Direct Manipulation and the Desktop metaphor aren't the best way to do everything. Remembering and typing is sometimes better than seeing and pointing. For example, if a user wants to open a file that is one of several hundred in a directory (folder), the system should let users type its name rather than forcing them to scroll through the directory trying to spot it so they can select it." (Johnson &c.: The Xerox "Star", a retrospective)
Btw if one askes for a transparant look at the content, isn't that asking for UNIX? You can't get more transparant than UNIX.
we have a standard interface, but are we happy with it?
Another way to approach this question is to maintain that we have in a certain sense a standard interface. Does anyone of us use a computer that is not using a WIMP-interface? Who of us does not use a GUI? A GUI is already a lot more specific that a book. (A book after all is nothing else than a stack of folded paper bound together). Going even farther: we hardly use any program or interface that does not in some way resemble Windows - excepting old CLI-programs. (We do not have that many experimental interfaces that are used widely). So seen from this perspective the problem might not be that we do not have a standard user interface, the problem is rather that we do have a standard user interface, one that has already set the conventions for interaction with computerdata, one that is in some ways highly restrictive. (It was designed for office use - and not all of us work in offices).
Btw: I think we have this standard WIMP / GUI because it simply does some of the jobs in the best imaginable way. Displacing this standard could be already out of the question. (We might use voice and fingers pointing to a virtual screen as input devices in the future).
Using other interfaces will always mean that we have to spend some time trying to figure out how it works. The very beautiful 'interactive book' made by David Small - shown on the last Documenta - could figure as an example. After reading the instructions and after browsing through it for ten minutes the viewers still had no idea how to work with it.
still another critique
Still another critique could be that it's not such a good idea to compare books to computers. The book is a far too small category compared with computers. The comparision should run between computers and all sort of paper-based storage- and interaction-devices (database-cards, leaflets, notebooks, etc.)
In the end it's a question of how we visualize information. The question of how data become information, and how that is turned into knowledge. How much of that is 'quasi universal', - the part that has to be standardized - and how much is culture-specific, how much depends on context, how much on the envisioned users?
the real issue at stake
Talking about standards: having standards for software and protocols is of much larger importance. Standards that make sure we will continue to be able to read and exchange information - without selling ourselves and our information de facto to a corporation. I cannot stress this enough: we have to keep fighting for the use of open standards: TCP/IP, HTML, open source tools. This is way more important than a kind of standard user interface. As long as we use open standards, we will be able to read data that is stored digitally, and read it using different applications, with different user interfaces. And we have to remember: before a human can read digitally stored data, a computer has to be able to read it. Emulators are as we know a good solution - but how much will "they" charge us for the use of an official emulator?
All this, by the way, does not imply that we shouldn't have to do our best in creating good, well-readable, navigable books, websites, database-interfaces and the like. Some of the aspects of what constitutes good design are easily intuited (and sometimes difficult to implement) - like readibility of letterforms - some are highly dependent on context (education versus entertainment), some are highly dependent on the culture of the users (unix-users versus GUI-kids). Also I think that the idea to look at the experience of users as a key of what constitutes a good interface (n no user is the same), is a good one. It means a.o. looking at how users accomplish 'tasks', how they go about 'getting at information', or 'attaining knowledge'.
And here I could follow with a critique of the existing paradigma of usability, but I will rest my case for now.
A last remark: in the world of information design there seems to be a shift from the design of good interfaces for applications, single websites and/or CD-roms, towards an interest in designing interfaces that enable users to access their own information, share it with others, link it, knit data together, regardless of platform, technical protocol etc., so regardless whether the data is an e-mail, a word document, a jpeg, a piece from a cd, regardless if you'd like to view it on a PC or a notepad or a mobile phone. If you are in the business of making information (archives) accessible, this is something to keep in mind. (But again, in the end this leads back to the stressing of the importance of open protocols, and a strict separation of content and design.)
some literature | interesting links
Matthew Fuller: A Means of Mutation http://www.axia.demon.co.uk/mutation.html
Jeff Johnson and Teresa L. Roberts, William Verplank, David C. Smith, Charles Irby and Marian Beard, Kevin Mackey: The Xerox "Star": A Retrospective http://www.geocities.com/SiliconValley/Office/7101/retrospect/
Paul Kahn and Krzysztof Lenk: Mapping Websites, 2001, Bis, Amsterdam
Susanna Pajares Tosca: A Pragmatics of Links http://jodi.ecs.soton.ac.uk/Articles/v01/i06/Pajares/
Neal Stephenson: In the Beginning was the Command Line:
Edward R. Tufte: The Visual Display of Quanitatative Information, 1983, Graphics Press, Cheshire id: Envisioning Information, 1990, Graphics Press, Cheshire
id: Visual Explanations, 1997, Graphics Press, Cheshire
Colin Ware: Information Visualization, Perception for Design, 2000, Academic Press, San Diego &c.
some rights reserved