Vol. II, No. 3

November, 1989

A CADD model may be likened to a scale model of a complex object -- except that the CADD model cannot be touched, but only seen from any angle or viewing distance. An accurate scale model may be disassembled and each piece examined separately, examined along with any number of other pieces, or examined with all the pieces; similarly, a CADD model consists of a number of individual parts which may be examined separately, with any number of other parts, or with all the parts. When combined, the parts of the CADD model -- called layers (the term has no relationship to an excavation layer) -- make up the complete model; separately each layer may contain any portion of the overall model. The selection of material for a given layer is at the discretion of the draftsman and may be entirely arbitrary. However, the creator of a model invariably uses a scheme of some kind to determine what is placed on each layer and to name his layers so that he can better remember the contents of each (there may be hundreds of layers for a complex model). Names must be meaningful, but they will be difficult to use if they are too long. Therefore, in any complex layering scheme, there must be a compromise between length and obviousness; an extremely complex scheme will inevitably become almost totally abstract.

A good naming convention must also make it possible to search for groups of layers easily (all of a given date range or all pertaining to a certain structure, for example). That, after all, is one of the features of CADD which makes it most helpful for scholars. Computer searching schemes, as used by most CADD programs, permit users to search for layers efficiently only if each letter in the layer name designates its information according to its position as well as the letter itself; for instance, an architect's building plan might have two positions at the start of each layer name so that a two-digit number would begin each name, with the number indicating the floor of the building to which the layer pertains, and the next letter might indicate that the layer contained walls, plumbing plans, heating and air conditioning schemes, or some other subsystem. With such a scheme, one could select only those layers relating to any floor or only the heating and air conditioning plans, etc., but to provide such assistance, a naming system must be carefully crafted.

Layer names which are both reasonably short and properly constructed will provide important benefits to any CADD user, but to meet the needs of the archaeological community a standard naming convention is more critical, indeed it is essential, if scholars are to be able to examine models with which they are unfamiliar. CSA is undertaking an effort to establish a layer-naming convention which can be used widely. Please send your thoughts, comments, or suggestions about this to Dr. Eiteljorg at CSA. (This will be an important part of the workshop meeting at the AIA annual meeting in Boston.)

To assist in crystallizing your thoughts about layer names, the following scheme may be taken as a "stalking horse." It is intended to provide a thorough scheme for naming and grouping layers and for some flexibility to meet the needs of the individual scholar, but its primary purpose is to attract comments, suggestions, and criticisms so that a better and more comprehensive scheme may emerge.

The length of the names of the layers has been kept to a minimum -- 9 characters plus dates or excavation trench and stratum; this makes the system necessarily abstract, with single letters carrying complex meanings. The suggested scheme permits the letters of the alphabet, the numerals 1 through 9 (not 0), and the dash or minus sign (not the underline); in the date positions only the numerals 0 through 9 and the minus sign may be used. The specifics of the proposal are as follows:

• The first character would be used to indicate the kind of drawing, as opposed to the kind of archaeological material, on the layer. For instance, it is more convenient for viewing drawings if dimensions, notes, labels, and cross-hatching are placed on separate layers so they do not intrude when they are not needed. It is also desirable to have both 2-d plans and 3-d models (see "Two-Dimensional Drawings from Three-Dimensional Models" in this issue, but not currently on the web) of much of the material, and the 2-d and 3-d layers should be separate. Thus, layers beginning with the letter M would be those with 3-d models, those beginning with P would have simpler 2-d plan views, those beginning with D would have only dimensions, and so on.
• The second character would be used to indicate whether the material drawn was a movable artefact, structural/architectural material, natural bedrock, bedrock trimmed to be structural, undisturbed earth, etc.
• The third character would be used to indicate whether the feature/object had been disturbed, was a random find, was found in situ at the point of last use, was found in situ but moved from its last point of use (with different letters to indicate the cause of the change, e.g., subsequent natural shifts or destructive agents, natural or human, etc.). Numerals used in this place would indicate hypothetical material, permitting as many as nine different reconstructions of any specific feature.

The fourth and fifth, sixth and seventh, and eighth and ninth characters would be used in pairs to provide combined general (first character of the pair) and specific (second character) information about three categories -- the area of the find, its usage, and the material.

• Thus, the fourth character would designate the general type of the finding area, such as public area, domestic area, military area, work/manufacturing area, and so on. (A "not applicable" choice would be available for this and all positions, but it would be possible to have a clearly not in situ random find -- according to the letter in position three -- which was nonetheless located in a public area.) The fifth character would provide more specific information, based upon that supplied by character four; for instance, a feature/object located in a work/manufacturing area could be further located as coming from an open square, hall, storage room, road, kiln, burial, etc. The fifth character could also be used to indicate that the attribution in the previous character is uncertain.
• The sixth and seventh characters would indicate usage, but they would carry different meanings, depending upon whether the material is a movable artefact or not. The first of the two letters would signify a broad category -- e.g., wall, column, or roof for a structure; weapon, cooking utensil, or tool for an artefact. The following character would again narrow the range of the first -e.g., base, drum, or capital for column; bow, knife, or point for weapon.
• The eighth and ninth characters would be used to indicate the material of the feature/object. Once again, a general category would be given by the first of the two characters and a more specific one by the second. For example, the eighth character might designate human bone, precious metal, building stone, etc., and the ninth would specify adult male, adult female, or male child; silver, gold, or electrum; marble, poros, or limestone.

An advantage of the two-character pairs is that, if necessary, an individual scholar could adapt the second character of any (or all) of the pairs to the needs of a specific site or building without fatally compromising the entire scheme. For instance, an excavator in Rome might wish to have a very specific list of building stones for use in category nine to specify individual quarries. He could still use the basic scheme; although the specific choices for building materials would be "non-standard," they could be explained in information accompanying the data file. Scholars would not need to know the specific materials if they only wanted to see the structure, since they could call up all the layers of building stone without regard to the specific stone used. Scholars interested in the quarry information, on the other hand, could still have access to it.

• The remaining characters in the layer name would indicate the date of the material or the excavation trench and stratum. For giving dates, the earliest date which can be applied to the material in a given layer should be entered after the ninth character, with five characters required (using a minus sign for dates B.C. and an extra leading zero for dates A.D. because the plus sign is not permitted). Then an ending date for the material (destruction date or date of last use) should be added (again using 5 characters, including a minus sign or a leading zero). The dates may be specific to the year, decade, quarter century, century, etc., but must be expressed with numerals (e.g., if the date is fifth century B.C., then the beginning date would be -0499, and the ending date would be -0400). For an excavation these ten characters (or fewer, if ten were excessive) would be used for trench and strata designation or lot number, as specified to fit the scheme of the excavators, but, at the conclusion of the study period, a second data base would be created, with dates instead of the trench and stratum information and with material from various trenches combined in many instances (though the original data base would be preserved).

It should be noted some material may exist on more than one layer. For instance, a triglyph found in a disturbed area would be drawn in its findspot on a layer identified by its findspot, usage, material, the dates applied to its excavation level, and so on. It could also be drawn on a layer identified by the date of the structure from which it is thought to have come, its use in that structure, and its material -- and located where it would originally have been. That layer would, however, be identified as hypothetical.

For other Newsletter articles concerning the applications of CAD in archaeology and architectural history or the "CSA CAD Layer Naming Convention," consult the Subject index.

Table of Contents for the November, 1989 issue of the CSA Newsletter (Vol. 2, no. 3)

Table of Contents for all CSA Newsletter issues on the Web