One of my most difficult jobs is the evaluation of new CAD software. It takes a long time to become familiar with a CAD program, and it would be unfair to compare a program I have worked with for days or weeks to one I have worked with for years. Nonetheless, new programs - or improved versions of old ones - are constantly coming to market, and many scholars yearn for a program simpler and more intuitive than AutoCAD. (1) Because it is so difficult to learn complex new programs, I have developed a habit of starting with CAD programs by looking for the features I believe to be essential but likely candidates for omission. I do not worry about whether a program seems difficult to use, just whether it can do the job. If it can, I will continue to work with the program, try to get used to its idiosyncrasies, and judge it more fully. If not, I contact the technical support people to make certain that I have, indeed, identified serious flaws from my perspective. If those flaws are truly devastating, then I do not spend additional time with the program; to do so would serve no purpose. Such was the case with MicroStation, reviewed in the last issue of the Newsletter. Despite its wonderful working environment, it could not be coerced into creating certain kinds of surfaces. Nor was its layering system flexible enough for scholarly use.
In making these kinds of appraisals, I try to avoid commenting about the way users interact with the programs, because that is something that can seem very strange to a new user but familiar and even desirable to one who has used the program for years. (In my review of MicroStation I did comment on the working environment, because it was so good that it deserved special praise.)
MiniCad is a CAD program I just put through this kind of trial. Recently introduced in a Windows version, for Windows 95 and Windows NT, MiniCAD (in its current version, called MiniCad 6) has a long history as a Macintosh program but has not previously been available in a PC version. I was eager to try it, because a good CAD program with equivalent MAC and PC versions would be very desirable. I tested the Windows NT version on our new Pentium Pro machine.
Despite my high hopes, I concluded, contrary to what Mr. Paul Zimmerman concluded in his review (see "MiniCad 6 - Another Review" in this issue), that the program lacks essential features as it stands. However, I also concluded that the possibilities of improvement are rather good. So, while I do not recommend the program and will not continue to use it, I will keep an eye on future upgrades in the hope that critical improvements will be made.
The quickest way to check a CAD program for me is to see how it handles complex surfaces such as those I have dealt with in Pompeii. These are non-planar surfaces on walls that are not vertical. The boundaries of the individual surfaces can be surveyed; so creating them should be fairly straight-forward. But this is often an impossible task for CAD programs. (Please note that using surfaces in a CAD model is crucial. Otherwise, neither hidden-line drawings nor renderings can be created.) MiniCad failed this test. Though I could draw a surface with various processes, including rotating a shape around an axis and extruding, I could not make surfaces from individual points in space by simply using the x-, y-, and z-coordinates of those points to identify the surface boundaries; there was no way to input the coordinates. Therefore, starting with survey data, as we do, we could not use MiniCad to model the surfaces we measured in the field. Indeed, for all practical purposes, no surfaces found in archaeological work could be modeled, since completely planar surfaces are so rarely encountered. (It is possible to create entities similar to surfaces, MiniCad polygons, from individual points either through a two-step, enter-and-edit process, two slightly different versions of which are described below, or by writing programs in the language supported by MiniCad, called MiniPascal. However, surfaces are basic to 3D modeling; I should be able to create surfaces directly. Furthermore, MiniCad polygons appear as surfaces in MiniCad but only as connected lines - not surfaces - if the model is exported using the standard CAD export format, .dxf.)(2)
I looked first for the problem with surfaces, because it was one I had expected. Indeed, a similar problem was found in MicroStation. I did not expect to find it to be very difficult to make a line or polygon in MiniCad. However, using coordinates typed at the keyboard, I could draw only simple, horizontal lines or polygons, with all points having an elevation of zero. So I could not draw a survey line that, for instance, records the perimeter of a building site or excavation trench by entering the known coordinates from the keyboard. I could work around this problem in two ways. I could draw a polygon by typing the x- and y-coordinates of each vertex; an elevation of zero is automatic for all vertices. Then I could convert the polygon to a 3D polygon, and edit the z-value of each vertex. Or I could use the mouse to make a 3D polygon approximately the shape needed and then edit each vertex. Only this enter-and-edit process will work to draw 3D entities with known coordinates. It seems very strange, but, indeed, keyboard entry is possible for coordinates for the so-called 2D polygon, but not 3D polygons. Is this problem a question of how the system works, or is it a missing feature? I would call this a missing tool, rather than a missing feature. I could, indeed, make some 3D figures, using typed coordinates; but I could only do so in an indirect way, by creating a figure and then editing it. However, this is such a basic tool, in my view, that it rises virtually to the level of a missing feature.
Other problems with MiniCad included the use of layers in 3D and the use of layers as data segments. Readers of the Newsletter are by now familiar with the idea of layers in CAD models, I hope. Each drawing layer is like a transparent overlay, holding parts of the drawing which, for any reason whatever, are separated from other parts. Layers in MiniCad are either two-dimensional or three-dimensional layers, and each 3D layer in MiniCad has its own elevation, an annoyance but not a serious problem. However, using 3D layers together requires that they be explicitly linked, a major annoyance. More important, neither the layers nor the classes of objects (more nearly like a layer in MiniCad) can be manipulated with simple search procedures. But such manipulation is crucial to the efficient use of CAD models for either archaeological or architectural history work.
Since some layers are two-dimensional and some three-dimensional - and some tools are for two-dimensional work and others for three-dimensional work - I found it very difficult to work with the tools properly. Though that would surely change with time, the idea that there are two- and three-dimensional parts of the model that cannot be mixed seems to me both unnecessary and confusing.
In my experience with other CAD programs, I have become accustomed to the notion that the model is always scaled at one-to-one; all dimensions are entered at full scale. Only when making a paper copy must one specify a scale. MiniCad, however, obliged me to specify the scale of the drawing when I began, as if I were working on paper. That is not a problem with a simple plan, but it complicates the use of 3D models, since scale in a 3D view is useless. Furthermore, it complicates the process of making a paper drawing, since the paper drawing is, by default, produced at the drawing scale. Any other drawing scale must be stated in terms of the scale used on screen, not in real-world terms. That is, had I made a model using a scale of 1:1000, I would then have to define the output as enlarged by a factor of ten in order to create a 1:100 drawing. This is an annoying illustration of the program's roots, but it is not a disabling problem. (The use of drawing on paper as an operating metaphor also limited my ability to produce a drawing of specific parts of the model. One expects to be able to make a window of some sort around an area to be plotted for a particular purpose, but I found no provision for that in MiniCad. I could limit the drawing to a specific page and paper size but not a specific selection window.)
I am accustomed to thinking about CAD models as data sources first and visual presentations second. For that reason, I look quickly at new programs to see how users can retrieve information. MiniCad has a very good information box for basic facts about individual objects, and it is possible to obtain dimensions or angles from the model. In addition, the query tool allowed me to obtain the distance from point a to point b - and to continue on, getting a total distance from point a to b to c . . . - a very handy feature.
MiniCad has a tightly integrated spreadsheet that can work like a connected database, and there is a programming language included. I have not devoted the time required to learn to use either of those features.
Mr. Zimmerman, in his review, points out the possibility of linking surveying instruments to MiniCad so that some figures can be generated automatically, and he and his colleagues have been able to do that very effectively with the MiniCad programming language. Though the resulting system would not make it possible to model irregular surfaces, it can overcome the problems of creating other 3D figures from surveyed coordinates. However, these are not the capabilities of MiniCad but of MiniCad with added software and a total station. Most CAD programs could match the automation with added off-the-shelf software and no programming.
Even using MiniCad with the added features described, I would not be able to create new 3D figures by entering their coordinates directly from the keyboard. I could only do that through the arduous process outlined above. Thus, any data not automatically transferred would be problematic to enter. (Since it is possible to use keyboard entry for two-dimensional figures, I suspect that the program may change for the better. Such a change might not make it possible to create proper surfaces, but at least standard lines and polygons could be entered from the keyboard with minimal modification of the program. Therefore, I will watch for upgrades to come.)
My assessment of MiniCad as it now stands is a strongly negative one. I do not believe it can be used to create an accurate 3D model of the surfaces that comprise archaeological material or standing architecture. No benefits, in my opinion, can overcome that shortcoming. Since, in addition, there are some important missing tools and unnecessary confusion between two-dimensional and three-dimensional tools and processes, I cannot recommend the program. For a different view, see the following review by Paul Zimmerman.
Next Article: Publishing on the Web - Making Decisions for the Future
Table of Contents for the Nov, 1996 issue of the CSA Newsletter (Vol. 9, no. 3)
Table of Contents for all CSA Newsletter issues on the Web
Return to CSA Home Page
1 Using AutoCAD for a time, however, shows clearly that intuitive and efficient are not necessarily the same. While it is not an easy program to use at the outset, it is a very efficient tool in the hands of one familiar with its syntax. Precisely because the commands may be typed, the experienced user can work very efficiently by directly typing commands instead of searching for them in a maze of icons. Return to body of text at note 1.
2 Making such surfaces with AutoCAD is very time-consuming, and, for complex surfaces, I have resorted to programming to make the job quicker and easier, but it is possible to create complex surfaces without programming, and the surfaces thus created are exported correctly. Return to body of text at note 2.