[Return to Home Page]
Interfaces - Future Interactions
This essay was originally written for "Workbench" the newsletter of the Amiga Users Group in 1996.
PART 1 compares the Amiga User Interface (Workbench) with Windows 3.1. (This seemed pretty much irrelevant when considering a reprint in "Reality Module No.5" in August 1998, and I published PART 2 only.)
Now, in 2002, I am publishing the entire article online for historical reasons.
Consider an office-worker on their way to work. They do not think to themselves: "I'm first going to load WordPerfect, and then I'm going to load Microsoft Excel." Rather their thoughts go : "I've got to finish off that letter to Colonel Sanders, then I've got to enter the July figures into the sales spreadsheet." People are project-orientated, not software-tool orientated.
Windows is a menu-system for launching programs. (The group-menu icons bear no resemblance to the actual arrangement of files on the harddisk. They have been selected and arranged arbitrarily.)
Windows 3.1 has three ways of presenting harddisk contents - 1). The group-menus (these show selected applications only), 2). File manager - this shows the actual structure of directories and files. (It works okay but is ugly), 3). The file requester in the application you are using. (In my opinion this is very poorly designed. Drives, directories and files are shown in separate windows - effectively destroying the organisational structure which is the great strength of a hierarchical filing system. Furthermore - too few files are presented at once, and scrolling is awkward.)
If Windows 3.1 could be seen as a mutated menu-system, what is Workbench? Workbench is a mutated directory utility! Its strengths over Windows 3.1 are: 1). You don't need to launch a program before you can do file management functions, 2). It is WYSIWYG - file and directory locations are analogous to actual file hierarchy, 3). Changes made are shown immediately (you don't need to fiddle with your files and then fiddle with your GUI), 4). Manual installation of software is much easier - drag file icons to appropriate directories, perhaps write an assign statement, then reset your machine. (With Windows you could use file manager to drag files to appropriate locations, then edit your AUTOEXEC.BAT, then run the PIF editor and prepare a Program Information File to tell Windows where to find your file, then attach an icon to your Program Information File, then name your icon, then reset your machine!)
When I am using Windows I always feel like I am directing a robot to carry out tasks for me, with Workbench I feel like I am carrying out the tasks myself directly. Workbench gives me a greater feeling of control.
The Windows 3.1 environment is at its most absurd when dealing with floppy-disks and CD-ROMs. When you insert a floppy in a PC the disk isn't read automatically and a disk icon does not appear automatically on your screen. PCs have to be told to read floppies, and changing default drives is always more awkward than it should be.
CD-ROMs are treated like program files in Windows, in that they get group-menu icons assigned to them. The ridiculous thing is that these icons remain whether the relevant CD-ROM disk is loaded or not! The only way to find out whether a particular CD-ROM disk is inserted (when there are several CD-ROMs available), is to click on the icon and hope you don't get a "file not found" message, or to open the CD-ROM drawer and take a look at the disk! I've used CD-ROM databases in libraries I've worked in, and the way the PC handles them has struck me as a sick joke.
What is the most "natural" way to interact with your computer files? With "real" files you can have a series of labelled folders on a shelf and scan them with your eyes. Then you could grab the one you want and turn to the contents page. (With a computer you'd make use of the hierarchy of directories and subdirectories.) If you were starting a new project you'd decide which folder it best fits into, update the contents listing of that folder, and then put in a fresh sheet of paper.
Windows sabotages this by forcing you to choose your "tools" before you can choose your projects.
The Workbench approach is better because you can choose a file to update by double-clicking on its icon - and the ".info" file will grab the appropriate "tool" for you. (But Workbench is not perfect, opening these windows takes time. This explains why many Amiga-users like high-speed directory utilities like Directory Opus.)
I enjoy having the projects I'm working on shown as a neat array of icons in a Workbench window. It makes them seem more significant, and I like being able to load directly the ones I want to update.
I often think about what computing will be like in the future. I imagine a time when all software will be modular and interacting - analogous to the components of a hi-fi system. When starting a project on your computer would involve the operating-system launching dozens or hundreds of program threads, all interacting, sending messages to each other, and launching other program threads which flicker in and out of RAM. All fitting together, all working together.
You do not begin a new project by clicking on a program icon - a la Windows. (Indeed your software modules might have a harddrive partition all to themselves, and the only time you'd look at them would be when you were updating some of them. Most of your onscreen icons would be directory-icons for classes of projects.)
Instead you select from a pop-up menu or an icon "Create new project" - and then a blank page appears on your high-resolution screen.
Only then would you need to consider the type of data you want to work with. You may choose "text" from an icon or menu, and word-processor icons and menus would appear around your page. You would type away. Then you might outline an area of your page, and select "spreadsheet" from a menu - and while your cursor is in that outlined area your word-processor icons and menus are replaced by spreadsheet ones. (The page would be rescalable and while in that special area you would be using a spreadsheet). Other areas of the page might be defined by you as graphics, video, musical notation (with orchestral accompaniment if you wish), scalable windows into other pages, or live links to Internet sites. The areas with differing definitions would be rescalable and could be any shape or size you wish - *and* they could overlap! (Thus a spreadsheet could have a pastel graphic in the background, or a page of text could be animated.)
Much of what I have described above is being prototyped in computer science laboratories now. It is obvious that today's most popular user-interface is not suited to project-orientated computing, and software will need a radical redesign. (From "chunky" programs with few components, to programs with a host of small interacting components which are designed to work in a multi- threaded, multi-tasking environment. In olden days programs were independent entities, now they will have to communicate with each other.)
This type of computing-environment is possible on the Amiga - we have the components already: multi-threading, true multi-tasking, shared function- libraries, ARexx [for inter-process communication and control], etc. - all we need is appropriately re-engineered software.
Thus I laugh when I hear people describing today's personal computers as being "state of the art." We are *not* approaching the end of a computer revolution - a 2010 when everyone will sit before their Intel multimedia PC and feel a deep and abiding contentment! No - every aspect of the computer will be re-engineered anew, to operate efficiently in a new way. I see where we are now - but I am very much aware of just how far we still have to go with our technology.
Note added when PART 2 was reprinted in "Reality Module No.5" in August 1998.
*Of course now (1998) I am not so sure. Microsoft despite its protestations is really stifling innovation - we are getting an ancient operating-system retro-fitted every few years, and I wonder how much longer this can go on before the whole thing becomes unworkable. (You don't need an efficient or even a particularly good product when you have got the lion's share of the marketplace - all you need is some gee-wizz gewgaws in your new release to excite the geeks, and a major advertising blitz.)
The problem of proprietory formats is discussed in All The Dead Data.
Copyright © 1996 & 1998 by Michael F. Green. All rights reserved.
Last Updated: 16 March 2003