IAN'S QUEST

by Ian Chadwick

Ian Chadwick is a technical writer and the author of the 8-bit classic, "Mapping the Atari." Aside from writing, his other hobbies include zymurgy, paleontology and Napoleonic studies. He lives in Toronto with his wife, Susan, their six cats and Kepler, the dog.

Computer Aided Design (CAD) programs hold a particular fascination for me. I've always enjoyed model building and design, despite my absolute lack of physical skills in that area. Fortunately, the computer has provided me a medium through which I can work without getting glue or ink all over my fingers.

The ability to design objects and layout plans on the computer is very appealing. I have spent hours creating imaginary objects or modelling real ones such as Stonehenge or a street map of my neighborhood. For the Stonehenge model, I did considerable research about the position and design of the stones, then had to use my wits to create the model in the limited resources of the program.

CAD isn't a definition as much as a general catch-all for a miscellany of program types. There are three basic categories of CAD programs: modelling (also called Computer-Aided Modelling [CAM]; this includes CAD 3D, Master CAD and DynaCAD), drafting (Drafix, First CADD, GFA Draft Plus) and specialty (Computer-Aided Engineering [CAE]; printed circuit design is one such example). I'll be reviewing several of these packages in future issues of ST-Log, but I thought I'd use this column to introduce the topic.

CAD is an application with considerable impact on certain professions, especially engineers and architects. It is for them what the pocket calculator is for scientists: A device that increases productivity, reduces mechanical work time and releases the user from the more mundane aspects of the job in order to allow him or her to concentrate on the creative and theoretical parts of a task. People with good creative and design sense are no longer limited by mediocre artistic talent, because CAD programs can better substitute for paper and pencil.

The ease with which you can redraw and alter designs in a CAD program as compared with doing so on the artist's table is impressive, as is the speed with which it is done. Computer users have come to expect such advantages—word processors are a similar example where the computer has changed the whole nature of writing through the ease and speed of use. CAD programs will have an effect on the craft of drafting just like electronic typesetting machines had on typography. Mechanical drafting is a skill on the way to extinction, the work now being done by the designers themselves on the computer.

Tom Weeks, an architect friend of mine, told me how his firm employs many computers and CAD programs in their design and layout process. He himself, impressed with the lineup of available software, purchased a Mega-2 ST with which to work at home. You'll find some of his comments in my program reviews. His professional point of view, as well as his familiarity with a variety of similar PC-based programs should provide a better insight into this software than I alone could offer.

A recent article in Computer Dealer News summarized a report entitled "CAD/CAM, CAE Users: Current Applications and Future Directions," which surveyed the IBM PC CAD/CAM market. The survey found that 89% of firms using CAD applications did so for two-dimensional mechanical drafting. That means 3-D modelling is not a priority for professional users, a result that surprised me considerably.

I would have thought that modelling had a higher priority of use. After all, when a client wants a house designed, he or she is presented with a series of 2-D illustrations and blueprints. Perspective drawings are also made, which attempt to convey the "depth" missing in the 2-D drawings. In some cases, especially when office or public buildings are concerned, a model may be built after the basic plans are approved.

A good modelling program can leap ahead of the paper process. The client can sit down in front of the screen and "see" the house from any angle, rotate the view, zoom in to examine detail, look "inside" to determine layout. If properly designed, the roof could be a cohesive object group that can be removed or temporarily hidden. Perhaps each floor could be grouped so that, from the perspective of looking straight down, the successive removal of roof and floor levels would provide a sequential blueprint effect for the viewer, showing floor plans clearly. Finally, using those programs so capable, an animated tour of the house could be designed, taking the client around and through without intervention.

I suspect that, despite the obvious value of such a presentation, most architects and professional users are wedded to their traditional display methods and tools, and therefore haven't graduated to using the computer to its full potential. Newer generations, schooled in these automated functions at the university level, will bring their profession (kicking and screaming, no doubt) into the modern era.

Drafting applications are not to be sneered at: They are very important in the design process. Paint programs are not suitable as drafting applications, since they treat the screen as a canvas.

Once painted, the only way to change the canvas is by over-painting or block alterations, a brute-force method at best. Drawing and drafting programs are object oriented. This means that a line, collection of lines, an object (such as a polygon or circle) or object group are separate entities that can be moved, altered or erased individually, with no effect on other objects on the screen.

If this distinction isn't clear, here's an example. Draw a box on the screen in a paint program: a simple, four-sided rectangle Next, draw another so that it overlaps the first. Now try to resize the first without changing the dimensions of the second. It is possible, but only by erasing and redrawing lines or using block rescales and then patching up the differences manually. In a drawing/CAD program, you can simply select the first box and alter it, regardless of the proximity of the second.

Along with object orientation, CAD programs often offer "layering"—that is, the ability to draw on distinct levels. The advantage of this in an application such as electronic circuit boards is obvious.

CAD programs and paint programs also differ in their ideal output devices. Paint programs suit high-resolution dot matrix printers, preferably color. CAD programs suit plotters best. Both work quite well with those laser printers which combine both vector and dot-matrix print capabilities, although current 8.5 by 11-inch paper limitations certainly limit the output. Paint programs display best in low resolution, where the full range of color can be used. CAD programs work best in monochrome, where fineness of detail is at its greatest.

This isn't to denigrate paint programs. Draw/CAD programs usually lack such paint-oriented features as magnified pixel manipulation and sophisticated palette control and animation. Paint and draw programs function differently because they usually have different end-uses.

Personally, I find drawing freehand with a mouse a painful and inaccurate process. The tool is not adequate for the task, in part because the physical design of the mouse. First, the drawing "point" is not in the same location where I learned to draw with a pencil. Second, the mouse focal center, the ball, is misplaced—it is too low in the hand for precise work. The new Microsoft Mouse places the ball (and thus the drawing center) right below the fingertips and the buttons, a superior design, and one I trust others will soon follow. Finally, I look at the screen, not my hand, so I can't use simple hand-eye coordination in my drawing. I'd much rather use a drawing tablet, which acts and feels more like the paper and pen with which I learned. In CAD programs, the mouse is an excellent device (although I'd still prefer the more ergonomic MS Mouse) and a drawing tablet is often unnecessary, the mouse performing adequately.

Ideally, a combination of both types of programs would be best, but the sheer size and complexity of a combination program would preclude it from running either quickly or on smaller-memory machines. In the meantime, the better CAD programs at least permit saving the screen as a DEGAS or Neochrome format file so the image can be incorporated into illustrations. This is a critical link between the two types of programs. The other side of the coin is the ability to load a paint picture as the background onto which a model or drafting illustration may be drawn.

Because of the extreme differences between paint and CAD files, I think it necessary for the manuals to detail the file structures in order for third parties to design compatible programs around them or for hackers to be able to use the output in other applications. Of course, I believe that the file structure of every application should be explained in the program's manual, because secrecy over this sort of data only isolates the users from potential outside enhancements.

One of the main issues in CAD is precision scaling, not only in its ability to deal in absolute but also in relative terms. The programs need to be able to simulate various scales of reference accurately and efficiently. When designing a house, for example, you need to be able to build to exacting specifications, not simply create objects that "look" right. The same criteria applies to 2-D programs. A blueprint is a precise document and the program must be able to provide scaling and measurement functions to suit professional output.

The collateral aspect of the above is the ability to produce in listed form the technical details of an object or a group of objects. In the house example, if the client wants to know the size of a particular wall, I should be able to select that wall and print out the scale details: height and width. Better still, I should be able to list (and print) all of the technical specifications for all of the objects in the chosen measurement unit. In 3-D models, this includes other data such as distance from the universe center, depth and X/Y/Z coordinates. Without such information, modelling in two or three dimensions can be, at best, only an inexact and inaccurate process.

Two-dimensional and three-dimensional programs have considerable differences that go beyond the obvious dimensional differences. Objects in 3-D have different display modes (wireframe, wireframe with hidden lines removed, solid, solid with shading), variable lighting for solid displays, the complexities of solid object join and subtraction, variable perspective and movable point of view. There must also be one or more tools with which to create a 3-D object from a 2-D template; extrude and spin are the two most common tools. The ease and sophistication of these tools will be an issue I look at.

The reviews will begin next month and run for several months to follow, each one covering one program in detail. I hope you'll enjoy reading them.