CSA received a copy of the newest version of AutoCAD, Release 13, in May. We are very grateful to Autodesk, Inc., for the gift of the program and their continuing support. The new version of AutoCAD was taken to Pompeii for the '96 season; I used it while there and after my return. As I have become accustomed to this latest version of the program, I've also tried to learn what is new and different about it. The version I am using runs under Windows NT, which may make it slightly different from other versions, but differences should be few.
The first thing I noticed about the new version was the hardware lock, a small item that must be plugged into the printer port of a computer if the program is to run. Without the lock, the computer will run but not the program. As it turns out, we should have had a hardware lock on CSA's copy of Release 12, as academic buyers did, but we received a standard version of Release 12. So far as I know, only academic users and buyers outside the U.S. are obliged to use a hardware lock; regular purchasers in the U. S. can run the program without it. (The printer cable can be plugged into the hardware lock, and the printer will still work.)
This was my first use of a hardware lock on AutoCAD since its ill-fated introduction by Autodesk some years ago, but then all buyers were required to use the hardware lock. The results then were so disastrous that the devices were eliminated within months. Perhaps my experience with the first iteration of the hardware lock accounts for my aversion to hardware locks now; that first one simply did not work, and a great deal of time was lost trying to figure out why and then waiting for the arrival of a replacement. At any rate, I resent being required to use it (especially since, in the U. S., only those who have the discounted academic version must use one). It is simply one more thing to go wrong with a computer, and any added potential problem is one I don't need - especially in the field, not miles but thousands of miles from assistance.
The second thing I noticed was that the system was supplied on a huge number of floppy disks - 59 of them, not the CD I had requested. I suspect that the hardware lock version is not available on CD, but I'm not certain of that. Loading a program from that many floppies - even when not all are required is a silly waste of time, especially when some of the instructions are unclear - and when some of the printed instructions have been superseded by new instructions supplied on disk. That confusion of new and old instructions gave an amateurish feel to the package, not to mention some uncertainty about processes to follow.
Once the program was loaded, I found many minor improvements but relatively few enhancements of real importance. It does run much more quickly, though. The screen is refreshed in a fraction of the time previously required, and that is not a trivial improvement. There is also a new aerial view screen that permits the user to pan and zoom more quickly and efficiently. That is very helpful when working at high levels of magnification. (This was available in some versions of Release 12.)
The most important new feature for many of us is the ability to draw complex curved shapes more accurately. We can now draw NURBS (non-uniform rational b-splines) with AutoCAD. This will make it possible to draw more accurate pottery profiles, and I hope to have yet another experiment with pottery profiles ready for the next issue of the Newsletter. I will use this new feature to see if better - or easier-to-draw - profiles can be produced with it.
There are important new solid modeling commands and processes, but I have not tried to use them. Indeed, I have made relatively little use of the solid modeling features in prior versions; so a judgment about the new versions would be impossible. The changes in this area, though, are among the most important for AutoCAD, since the information stored about solids has changed along with some of the drawing processes.
Dealing with surfaces, on the other hand, is something I've spent a great deal of time doing, and I was disappointed to find Release 13 contains no new aids for making surfaces. The simple 3dface command remains the most flexible command available for making irregular surfaces. Although that command does not easily create surfaces with more than four corners, there are ways to use the command to do so and, indeed, to make very complex surfaces of many kinds.
Surfaces can also be constructed from four edges, but if the starting edges are irregular, AutoCAD does not do a very good job of matching the resulting surface to the original edges. Quadrilateral patches are created to fill the bounded area, and one must instruct the program to make many such patches in order to approximate the original boundaries. The result is still an approximation, and it does not produce a good hidden-line drawing, since the edges of all the patches will show.
There are routines for constructing regular surfaces and solids, but those are not the kinds of surfaces or solids we usually encounter when working on excavations or standing architecture. One designs regular surfaces, but one rarely finds them in the constructed world.
Since the new features of AutoCAD did not include helpful surfacing routines, we found it necessary in Pompeii to rely again on CSA's add-on routines (see "CSA LISP Routines Available at Web Site,"). Karim Hanna, who worked with me at Pompeii again this summer and concluded the season as the man responsible for surveying and modeling after my departure, helped to rectify a problem with one of the routines, and they worked very well, permitting us to make surfaces out of polygons, lines, or isolated points. (An added surfacing program from Autodesk, AutoSurf, is said to make the creation of complex surfaces much easier and more precise. The program is not too expensive - $250 for individual academic users - and we've been offered an evaluation copy; so I will try it out and report further in a coming issue of the Newsletter.)
New rendering capabilities in Release 13 add considerably to one's ability to produce interesting and understandable images. Lighting angles, intensity, and color can be adjusted, and the results are helpful. Materials for surfaces may also be chosen. I am normally more concerned with hidden line drawings, because they are more likely to be our ultimate output from work with a CAD program, but rendering offers many possibilities for presenting realistic images of ancient structures.
It is now possible to run multiple copies of AutoCAD under Windows NT. That permits a user to have multiple models on view at the same time. It seems strange to me that each model must have its own iteration of AutoCAD operating. It would seem simpler for the user (and reasonable for the programmers) to permit one program to handle multiple models, as a word processor, for instance, can have several open documents at the same time. Nonetheless, it is certainly a benefit to be able to work on more than one model at a time.
I did check the database connections in Release 13. It had not been trivial to figure out how to connect database information to drawing entities when that feature was added to the last version of AutoCAD, Release 12, but the instructions for Release 13 were, for me at least, virtually impenetrable. Fortunately, additional information was available on the Web; without it I would not have been able to figure out how to link a database to a model.
Having connected the data files to the drawing, however, I learned that the actual entities in the drawing - the individual blocks and icons that were linked to database information - had lost the links to that information when my Release 12 model of the older propylon was converted to Release 13. A search of the instructions yielded no help; so I sent email to someone in the educational department at Autodesk. With his assistance, I received, via email, further nearly impenetrable instructions about how to preserve the links between the drawing entities in the older propylon model and data in the FoxPro files. My first wasted hour trying to put those instructions to work resulted in a few false starts - the kind of glimmers of success that computers so often provide to prevent me from hurting myself - or the computer - while expressing my frustration. The next morning I managed to find my errors and to make the link; the data connections between the drawing and the FoxPro files was reestablished. It must be said, however, that this was an unnecessarily complex and arduous task. The programmers have obliged users to manipulate intermediate files with a text editor when they could have asked a very few questions in a dialog box and altered those files automatically. This is a user-hostile approach. Though I am told it may be improved in the very latest revision of the program, it illustrates the problem of trying to shield users from some of the complexity of modern software while simultaneously trying to speed the production line.
This process reminded me of AutoCAD's most glaring weakness - the absence of direct technical support. I was able to find help through the Web site and then the company, but most users have no access to the company when the manuals and the Web site fail. The local educational dealer charges - at the rate of $100 per hour - for technical assistance, even if you buy AutoCAD from them. This is simply not acceptable, in my view, especially when the manuals don't do the job.
The combination of the hardware lock and the poor support encouraged me to reconsider other alternatives. I spent some time working with the newest version of Microstation (see short review following this article), and I will be looking at MiniCad shortly. I remain an AutoCAD user, though, because it permits me to do things I can do with no other PC program that I know of, things I consider essential. I've found no other program that provides AutoCAD's flexibility for handling layers, and I've also found that, though not easy to use, AutoCAD's surfacing commands are very flexible and make it possible to model extremely complex surfaces. Indeed, particular problems encountered in Pompeii this summer were relatively easily handled with our add-on routines but could not be solved with Microstation. I would be far happier, though, without the hardware lock and with some reasonable technical support. I would also like to see a windowing system like Microstation's, as described in the next article, "Microstation 95."
For other Newsletter articles concerning the applications of CAD modelling in archaeology and architectural history consult the Subject index.
Next Article: Microstation 95
Table of Contents for the August, 1996 issue of the CSA Newsletter (Vol. 9, no. 2)
Table of Contents for all CSA Newsletter issues on the Web
Return to CSA Home Page