Arcà Andrea 2016. Digital Practices for the Study of the Great Rock in the Naquane National Park, Valcamonica, Italy: from Graphic Rendering to Figure Cataloguing

CAA2015 KEEP THE REVOLUTION GOING > » Proceedings of the 43rd Annual Conference on Computer Applications and Quantitative Methods In Archaeology edited by Stefano Cam pan a, Roberto Scopigno, Gabriella Carpentiero and Marianna Cirillo Volum es 1 and 2 rm rffilnJfciHlnll H IW ,UNIVERSITA Dl SIENA 1240 CAA2015 KEEP THE REVOLUTION GOINS » > P r o c e e d i n g s o f t h e 43 r d A n n u a l C o n f e r e n c e o n C o m p u t e r A p p lic a t io n s a n d Q u a n t i t a t i v e M e t h o d s in A r c h a e o l o g y edited by Stefano Campana, Roberto Scopigno, Gabriella Carpentiero and Marianna Cirillo Volume 1 A r c h a e o p r e s s a r c h a e o l o g y A r c h a e o p r e s s P u b l i s h i n g Lt d Gordon House 276 Banbury Road Oxford 0X2 7ED www.archaeopress.com CAA2015 ISBN 978 1 78491 337 3 ISBN 978 1 78491 338 0 (e-Pdf) © Archaeopress and the individual authors 2016 CAA2015 is availabe to download from Archaeopress Open Access site All rights reserved. No part of this book may be reproduced, or transmitted, in any form or by any means, electronic, mechanical, photocopying or otherwise, w ithout the prior written permission of the copyright owners. This book is available direct from Archaeopress or from ou r website www.archaeopress.com Digital Practices for the Study of the Great Rock in the Naquane National Park, Valcamonica, Italy: from Graphic Rendering to Figure Cataloguing Andrea Area [email protected] University of Pisa University, Dottorato in Scienze dell’Antichità e Archeologia; IIPP - Italian Institute of Prehistory and Protohistory Abstract: I f we intend rock art studies to be an archaeological science, the greatest possible accuracy should be obtained. In order to do so, computerprograms are indispensable; considering the needs o f the graphic and analytic workflows, a large set o f electro­ nic instruments should be intended as a fundamental toolboxfo r the careful and up-to-date rock art archaeologist. On the graphic side, the need to obtain a detailed rendering o f tracings may greatly benefitfrom vector drawing software. On the analytical side, the necessity o f managing figure classification and relations may lead to the compilation o f specific software. This is the case o f the study — undertaken by the author — o f the Great Rock o f Naquane in Valcamonica (Italy). Dedicated software was compiled andfrom this rankings, statistics, and a catalogue were produced in a few seconds in htmlformat, ready to be published online or printed. Keywords: Rock art, Valcamonica, Vector tracing, Database, Naquane Introduction and methodological concerns If we apply the same rules of archaeological science to rock art documentation, as we should do, including it as ‘rupestrian archaeology’ (Fossati et al. 1990), the study of figured rocks acquires the status of an archaeological excavation: a carved or painted rock panel corresponds to a site, its figures — which are ‘real’, although non-material, iconic objects — take the role of a collection of archaeological finds and the engraving phases equate to an archaeological stratigraphy. To ‘dig out’ such iconographic finds and layers, we need to ‘extract’ all finds, that is, figures, identifying and reproducing them graphically. Moreover, there is the need to ‘unearth’ the related superimpositions, if any exist and are readable, identifying, rendering, and enlightening them with an appropriate and effective symbolization. In this way, the establishment of an iconic stratigraphy is the key to obtaining the most detailed and correct relative chronological sequence; on the other hand, the comparison of iconic items and real dated objects, if feasible, is the way to establish an absolute chronology. With regard to engraved figures, we may specify that archaeometric measures are not practicable, at least not in the alpine environment;1 contrary to paintings, during the carving the engraved matter, which might be suitable for quantitative measures and for sampling and radiometric dating, has been removed, not added. To perform these archaeo-rupestrian tasks, a careful and in-depth recording is required. As with any other scientific archaeological documentation, it should be performed with the greatest possible accuracy. Regarding rock art studies, the long-lasting documentation experience (Anati 1966; 1 The micro-erosion dating technique, developed by R. Bedanrik and based on the analysis of quartz crystals, was tested in the alpine environment on the Rupe Magna of Grosio (Bednarik 2001). This technique, which has already obtained controversial results in Portugal for the Coa valley Palaeolithic engravings, has not been proposed again nor developed in the alpine area. 1974; Fossati et al. 1990) confirms that a quick and singular autoptic examination is not sufficient to identify figures and superimpositions — a great amount of time is almost always required to identify what is depicted in detail— and photographic recording (both 2D and 3D), although fundamental, may not guarantee clear reproductions, particularly if many figures are bundled together and/or the rock surface is worn out (Loendorf 2001: 65). There is the strong need to enhance contrast, extracting and abstracting figures from their bedrock and from nearby items if a scene is not involved. Above all, we need to reproduce the results by the most exact available and suitable graphic rendering (Fig. 1), appropriately symbolized by the means of flat projection, the selection of relevant traits, complexity reduction, design and visual effectiveness ready for communication and publication. All these characteristics make the production of this archaeological drawing somewhat analogous to the creation of cartographic maps, which are traced and drawn, rather than a simple reproduction of zenith photos. According to these premises, an archaeo-iconographic tracing and its rendering2 (Fossati and Area 2001; Area and Fossati 2006) constitute the indispensable base for any further chrono-interpretative study (Anati 1975), data sharing, and knowledge diffusion (Area et al. 2008). Some questions arise for surface-contact and objectivity concerns. Contact tracing (Fig. 2) is sometimes questioned, exposing worries about possible damage to rock surfaces and figures. This is surely not a problem where petroglyphs and strong rocks, which are hard and do not crumble, are concerned, provided that permission is granted by the authority in charge. Contact with the original item is largely preferable, being more detailed and assuring the achievement of a greater amount of visual, tactile, and material data. The problem of objectivity is more complex: tracings are often considered to be unreliable and not based on objective data. Regarding this 2 Methodological considerations and recent technical updating are treated in Maretta 2014 and Marretta et al. 2013. 1081 4 i = < ] £ CAA 2015 F i g . l . C o n t a c t t r a c i n g o f t h e G r e a t Ro c k o f N a q u a n e N a t i o n a l Pa r k , V a l c a m o n i c a ; s e c t o r P, SHEETS 23 AND 6, FAIR COPY OF THE RASTER VERSION (TRACING A A ). point, the problem is not related to this specific technique — any technique, manual or digital, even automatic, must comply at some point with human choices — but to the competence and experience of the tracing operator. So we may wonder, while performing this task, if an objective automatic scanner or an auto-recognizing software may (or will) go further than an accurate and experienced archaeologist; or, on the other hand, if and to what extent software tools may facilitate the experience of the archaeologist and improve his/her results. Nevertheless, it should be emphasized that the task of tracing figures and superimpositions is subject to unavoidable human choices in varying degrees; although as objective as possible, 1082 4 i = < m - A n d r e a A r c a : D ig i t a l Pr a c t ic e s f o r t h e St u d y o f t h e G r e a t Ro c k in t h e N a q u a n e N a t i o n a l Pa r k F i g . 2. A M O M E N T FROM THE CONTACT TRACING OF THE GREAT ROCK OF NAQUANE NATIONAL PARK, SIX SHEETS O U T OF A TOTAL OF 221, SECTOR P (PHOTO A A ). it takes the nature of an interpretive operation and its results are enriched not only by graphic elements, but also by ideas and solutions. The third question is related to the third dimension. 3D models, obtained by various methods such as stereo-photogrammetry, laser-scanning, or 3D photo-stitching and modelling, are gaining and will gain more and more importance in rock­ art recording. Their capability of avoiding contact and their objectivity are often flaunted as advantageous features. The two questions have already been treated above; the suggested solution would be to search for the appropriate development and reinforcement of such technical ‘experiments’, avoiding any specious devaluation of traditional and well-grounded methods, if still effective. The core of the problem is obviously the third dimension, that is, the recording of the z-axis, the axis of depth. While inconsistent for paintings, although obtaining the volumetric shaping of a painted rock wall is useful for conservation and exhibition purposes (Scott 2015) and sometimes also relevant for its meaning, it is clearly beneficial for engravings. That said, a distinction should be made between pecked figures and deeper ones such as cup-marks or basins. While a 3D recording of cup-marks and grooves is surely a good choice, the use of these techniques for shallow pecked figures may result in misleading and somewhat oversized recordings. First of all, it should be mentioned that such figures were intended and created by their makers as a series of combined dots. The same goes for their reproduction: the iconic item is the figure made by the texture of its points and by the resulting inner and outer contours. Both may be usefully recorded using a 2D contact tracing, where the point of a felt pen is driven by the eye of the tracer, facilitated by a multi-oriented grazing light over the engraved surface and by the tracer’s fingers, by feeling the presence of the pecked shallow dots on the rock through hand pressure, going ‘in’ and ‘after’ them. All this supports the points cited above, mainly that symbolization and complexity reduction for the final rendering, which is meant to show the iconic content, are better and more effectively performed by a 2D tracing than by a 3D model, not to mention the cases in which a manual sketch is performed on the monitor over a 3D model, clearly demonstrating the different purposes of the two systems. Clearly, the best choice would be a combination of the two, exactly like using photos to texture cloud points, in other words, simply merging a 2D tracing layer over a textured 3D model of the whole rock surface, as was recently done for the Chenal shelter in the Aosta Valley (Area et al. 2014). Although 25 years have passed since the first experiments of stereo-photogrammetry were precisely applied to the Great Rock of Naquane, it may be argued that, from the archaeological and iconographic point of view, no 1083 4 i = < ] £ CAA 2015 F i g . 3. T h e r ic h l y e n g r a v e d s u r f a c e o f t h e G r e a t Ro c k o f N a q u a n e , a m o n u m e n t o f t h e Ir o n A g e a n d Br o n z e A g e Va l c a m o n i c a r o c k a r t , s p e c t a c u l a r l y s m o o t h e d a n d s h a p e d b y t h e l a s t GLACIER; BEST RESULTS ACHIEVED W ITH THE OBLIQUE WINTER LIGHT (PHOTO A A ). publication about alpine engraved rocks with pecked figures (Valcamonica, Mt. Bego, and other areas) has ever gained and shown clearer and more defined iconic data from a 3D model than from a 2D-rendered tracing until now (Area et al. 2008: 13). 3D modelling, therefore, should be tested for small areas and should attempt to gain more definition, particularly for studying superimpositions where a different depth may show up a more ancient or recent figure and where 3D modelling may enhance such a difference and thereby be useful in reaching the aimed contrast. Another point to be considered is the recording of the conservation status: although a 2D recording is fully effective for the documentation of the various conditions of degradation, displaying their extension along the surface, a 3D model can record better the possible movement or detachment of small rock fragments or clasts. 1 NAQ1, the Great Rock of the Naquane Park FIG. 4. A DENSE AND VERY INTERESTING PART OF SECTOR G OF THE As indicated by its title, the scope of this paper is to present the methods — largely based on contact tracing and its graphic rendering — utilized for the documentation of the Great Rock, also known as Rock 1 (NAQ1; Fig. 3), of the Naquane National Park of Rock Engravings in Valcamonica (I), mainly focusing on software and digital involvement. The description of this archaeological monument shall be limited to stating that NAQ1, 50 m long and 4 m large, is one of the most important prehistoric figurative palimpsest of Valcamonica (De Marinis 1988; Fossati 1991) and ofthe entire alpine range. On its regular and smoothed rock surface more than 2000 prehistoric figures are stored, written, and archived on a Permian sandstone3 3 Its pétrographie nature — a very finely grained sandstone, pelite tending to siltite, bound by a silica cement — along with the glacial smoothing of its surface, was crucial to enable such easy and dense G r e a t Ro c k : B r o n z e A g e h u m a n f ig u r e s o v e r l a p p e d b y a n Ir o n A g e h u n t i n g s c e n e ( p h o t o A A ). surface, covering more than a 5 millennia-long chronological frame (from the first Copper Age to 1942). The largest numbers were carved during the final Bronze Age and the first Iron Age (10th-6th cent. BC; Fig. 4). Similar to all the Valcamonica (and Mt. Bego) rock art, figures were engraved by repeatedly hitting the surface with stone tools, most likely sharp quartz engraving activity. The same observation is to be extended to the entire Valcamonica and Mt. Bego rock art — in the petroglyphic complex of the Maritime Alps very similar pétrographie conditions are present — where by far the largest concentration of alpine engraved signs is present. 1084 4 i = < m - A n d r e a A r c a : D ig i t a l Pr a c t ic e s f o r t h e St u d y o f t h e G r e a t Ro c k in t h e N a q u a n e N a t i o n a l Pa r k sen. A sett. 6 sett, t sett.O sett. F sett, i sett. M sett. E sett. G sett. H sett, L F i g . 5. G r e a t Ro c k ( N A Q l ) o f N a q u a n e , s u b d i v i s i o n in s e c t o r s ( A - M ) o f t h e n o r t h p a r t , r o u g h l y h a l f o f t h e e n t ir e s u r f a c e (p h o to AA, pebbles, which produced a series of dots, a technique defined as ‘pecking’. Among the most important iconic items we only briefly cite here the so-called praying figures, dogs, looms (unique exemplars in the alpine rock art), towels, deer, warriors with important details such as spears, swords, shields, and plumed helmets, hunting scenes, duellists, and footprints. Similar to many others large engraved rocks of the alps, such as the Rupe Magna in Valtellina (Pace 1972; Area et al. 1995) — and unlike the Mt. Bego pecked panels, which were apparent — the NAQl figures were originally partially hidden under a light layer of lichen and moss. Its discovery, which occurred in the first months of 1932 (at the same time as the disclosure of the larger extent of the Valcamonica rock art) is claimed by two scholars: the archaeologist Raffello Battaglia — helped by his assistant of the archaeological superintendence Antonio Nicolussi and by a local informer Giacomo Bellicini — and the Piedmontese anthropologist and psychiatrist Giovanni Marro, with his local informer Giuseppe Amaracco. RafFaello Battaglia was the first to propose a well-founded archaeological hypothesis (Battaglia 1933), while Giovanni Marro noticed the absence of a unique frame of semantic relations and the presence of superimpositions (Marro 1932). The only complete study of the Great Rock of Naquane was published by Emmanuel Anati in 1960; thanks to the introduction of the analysis of different chronological layers, it showed a tracing executed in August 1957 over translucent wax paper — fully transparent plastic sheets were not available at the time and figure definition was quite simplified and some perspective distortion is noticeable (Anati 1960). Although the study seems to demonstrate that the published drawing was taken from photographs and not from a scale reduction of the original contact sheets; 876 figures were counted. From 1957 to the present, no updated tracing or a more detailed specific study or monograph has been published. r e w o r k e d ). 2 N A Q l, the digital graphic operational chain A new contact tracing of NAQl, performed by the author for his PhD research project at Pisa University,4was completed by producing 221 plastic sheets of 70 x 50 cm standard measure — along with five minor supplements and 34 connection strips — covering 65 m2 of engraved surface, subdivided into 21 sectors (Fig. 5). It took 50 inconsecutive days of on-site activity and about 150 days of graphic digital post-processing. As far as digital and software-related features are concerned, two main focus points may be outlined: graphics for the surface’s rendering and data analysis for its study. Inexpensive and easily obtainable digital tools were chosen, procured, or even specifically built as a convenience for rock art research, which rarely obtains sufficient, if any, funding — the new NAQl study itself has not been funded so far. Being a multi-phase and multi-tool procedure, digital graphic post-processing may be described as an operational chain. First of all, if large scanning machines (for paper sizes A3 to A0) are not available, original plastic sheets are optically reduced to 50% grey-scale, as colour photocopying is of no use, to fit the size of an A3 paper sheet. A 600 DPI B/W scan is then performed, equivalent to a 300 DPI scan of the unreduced sheet, thereby assuring that quality is not lost. Colours — during the manual contact tracing only red and blue are used, red for rock fractures and blue for collimation marks — are reinserted in the final rendering, as well as different grey or colour layers to symbolize chronological or typological differences. It should be specified that figures were originally traced only in black 4 At the request of the IIPP (Italian Institute of Prehistory and Protohistory) the study was approved and authorized by the Lombardy Archaeological Superintendence. I gratefully acknowledge Prof. Raffaele De Marinis, Dr Umberto Spigo, Dr Filippo Maria Gambari, Prof. Renata Grifoni, and Prof. Fabio Martini (Milan, Pisa, and Florence Universities, Lombardy Archaeological Superintendence). 1085 4 i = < ] £ CAA 2015 i- . t ■- ill ' F i g . 6. Ph a s e s o f t h e f in a l r e n d e r i n g ( v e c t o r v e r s i o n ) o f t h e N A Q l c o n t a c t t r a c i n g : i n n e r a n d o u t e r CONTOURS (ABOVE) AND FAKE COLOURS (BELOW ); TO GIVE AN IDEA, THE FIGURES OF THE TW O WARRIORS DUELLING W ITH THEIR SWORDS ARE COMPOSED OF 55 4 7 CURVE NODES (GRAPHICS A A ). to avoid confusion between phases and problems for the digitalizing process. A great amount of time, more than for the on-site activity, is needed to clean the scans digitally in order to obtain a fair raster copy produced sheet by sheet; these raster files are judged to be indispensable for conservation purposes, as in the computing future it will probably be easier to read pixel-based files than other formats such as vectors. To perform this cleaning task dirty spots are erased, hand-notes rewritten, frames and lines retraced, and figures separated; line retracing and figure detachment is essential to facilitate the subsequent steps of eliminating the occurrences of broken lines and improperly pasted objects. Thereafter, digital sheets for each engraved sector are merged as multiples layers, using the collimation marks to move and rotate them conveniently. Collimation marks and sheet frames are then deleted, thus obtaining a final draft, which is still in raster format with an image described by a dot matrix. Depending on the extension of the reproduced surface, the final raster file is very heavy and awkward to manage, especially if we want to maintain lossless quality. To cite an example, the raster merging of sector P of the Great Rock of Naquane, which, although 50% reduced digital sheets were used, is composed of 30 sheets, exceeding the 36.000 pixel limit and 1 Gb of uncompressed dimension on its largest side. Regarding the software involved, all the steps of this first part are easily performed — depending on CPU and RAM power — by raster photo-editors; since they are very well known worldwide, it is not necessary to cite them here. The second part of the digitalization process, far shorter and quicker, is devoted to solving the problems related to the weight of raster files and to their lossless rendering. Vector graphics, where images are not described by a dot matrix but by points, lines, and Beziers curves based on mathematical expressions, represent the best and most current choice. They 1086 4 i = < m - A n d r e a A r c a : D ig i t a l Pr a c t ic e s f o r t h e St u d y o f t h e G r e a t Ro c k in t h e N a q u a n e N a t i o n a l Pa r k F i g . 7. T h e f in a l r e n d e r i n g ( v e c t o r d r a w i n g ) o f t h e N A Q 1 s e c t o r C, 26 B r o n z e a n d Ir o n A g e f ig u r e s ; f ie l d - n o t e s a r e REWRITTEN, FIGURES ARE NUMBERED (TRACING AND GRAPHICS A A ). F i g . 8. T h e f i n a l r e n d e r i n g ( v e c t o r d r a w i n g , a b o v e ) a n d a g r a z i n g l i g h t p h o t o g r a p h ( b e l o w ) o f THE N A Q 1 SECTOR H, 72 BRONZE AND IRON AGE FIGURES; GREY-SCALE IS USED TO SHOW THE DIFFERENT SUPERIMPOSITION LEVELS, IN THIS CASE A M AXIM UM OF FOUR (TRACING, GRAPHICS, AND PHOTO A A ). 1087 4 i = < ] £ CAA 2015 were first applied to rock art tracings by the author in 1992 in the Susa Valley (Area 1999; 2000; 2009) and then, in the case of Valcamonica rock art, to the study of the Dos Cui engraved rock (Area 2005). Why vectors? To refer to the case of the NAQ1 tracings, vector rendering (Fig. 6) is essential for reducing file sizes up to 35 times; the sector G of the Great Rock, for example, is reduced from 380 Mb of the raster file to 10 Mb of the vector file. Most importantly, a lossless quality is gained, independently from the size scale, as each figure is a virtual object mathematically described by its outside and inside contour lines, the proportions of which do not vary with its dimensions. The stages of vectorization, provided a clean starting point is available, which is the fair copy cited above, are easily performed by auto-tracing software normally bundled with vector graphics editors, both commercial and freeware. These are illustration-oriented software packages, very useful for producing press-quality final renderings ready for publication or communication: the final outputs of tracings, plates, and tables can be managed in a graphically professional manner. In my opinion, vector graphics editors represent a fundamental tool for archaeologists, mainly for publication functions — for NAQ1, the CorelDraw suite was utilized. The auto-tracing process is performed by the software (CorelTrace) in a few seconds, even for very large tracings; a B/W bitmap file must be provided, as grey-scale or RGB images produce a very large and noisy set of fragmented graphic objects, according to grey or colour levels. In order to avoid the time­ consuming necessity of splitting lines, the best choice is to perform a first passage for figures, applying a contour auto­ tracing, and a second one for lines, choosing a central line auto­ tracing. The vector-based outputs are then merged as different layers in the final rendering. At this point useless parts are deleted, different sections of each figure are joined to produce an individual graphic object, figures are numbered, rock fractures are coloured, and their line-width standardized, while a general frame, headings, and scale measure are added (Fig. 7). Superimpositions, which are outlined by a thin white space among overlapped figures, may be further enlightened by using different colours or grey-scale levels (Fig. 8); grey-scale levels are recommended, since colours can be a problem for press output and for the unavoidable lack of a standard symbolization for different chronological phases or subject categories. At this point the job is completed; the final rendering may be used ‘as is’, some sections may be isolated, or each figure may be picked as a separate object and reassembled into typological or chronological tables or compared to other similar ones. All the described steps clearly constitute lengthy on-site and post-processing work. Even if someone claims that it is a time­ consuming process (Seidl et al. 2015), we need to recall that shortcuts are rarely the best choice in archaeological research and that the essential long time frame is necessary — with the best accuracy— to understand, study, and reproduce iconic palimpsests, which are often very tangled and hard to untie, for chronology and meaning. 3 NAQ1, data analysis software tools As soon as the graphic ‘digging’ has been completed and the iconic layers and figures have been extracted from the bedrock, there follows the need to study the collection of archaeological finds. Once more, a workflow is involved. The basis — the first and most important step — is the compilation of the catalogue of iconic objects and their relations. The figure record utilized for NAQ1 is the same as that specifically designed for Valcamonica rock art, which is also suitable for all alpine rock art; for the Great Rock of Naquane, classification opportunities have been enhanced. Its base structure originally derived from the Engraved Rocks Computer Recording Model, which was set up in 1989 by the Lombardy Archaeological Superintendence. Although the figure record is quite detailed and many alpha-numeric and Boolean fields are to be filled — dimensions, engraving technique, conservation level, and description — two main conditions should be outlined: taxonomic classification and identification of the network of relations. As for any archaeological find, a morphological and a chronological position must be assigned to each figure, along with a clear written description. For classification, and considering the need of simultaneously satisfying an overall and detailed analysis, a double taxonomical level is set. The ‘general category’ coding consists of a one-letter field; this currently includes 15 main figure groups such as anthropomorphs, zoomorphs, tools-weapons, inscriptions, cup-marks, and so on. For the northern part of NAQ1 — roughly half of the entire engraved surface — three categories prevail, of which two are not meaningful: zoomorphic figures (27.27%, coded as ‘B’), unclassifiable segments-lines-areas (25.16%) and dot groups or scattered pecking (24.76%). The ‘specific category’ coding is managed by a three-digit field, currently consisting of 273 specific categories; this more detailed taxonomic level is fully customizable for each particular rock art complex, site, or rock. The specific categories individuated for the Great Rock of Naquane (north sector) total 100; related rankings show the prevailing presence of warriors among the human figures and deer among the animal figures. A similar two-level structure is applied to the chronological classification: the ‘style’ coding consists of a one-digit field, currently expressed by 10 main archaeological periods such as Palaeolithic, Mesolithic, and so on. The ‘specific style’ coding is a five-letter field, currently listing 23 detailed archaeological periods such as Copper Age 1, Early Iron Age, and so on. To give an example, an A6 coded figure is an Iron Age anthropomorphic figure, while an A6VIA-30 coded item (Roman numeral VI) is a warrior of the first Iron Age, with a linear rectangular body, its legs placed as a triangle, and holding an upraised sword. It should be specified that the record compiler needs only to put in the code and not its description, except when modifying the code database. The second main condition, which concerns the relations network, is satisfied by recording, for each figure and if any exists, the associated iconic elements that have a semantic relation — useful for identifying scenes and giving suggestions for real or symbolic meanings — and the superimposed or underimposed elements, which are essential for revealing the sequence of the engraving phases. The current structure allows the inclusion of no more than three figures for each kind of relation, so a maximum of nine. Again for the north sector of NAQ1, 658 related figures, 250 superimposed, and 260 underimposed, have been recorded; 50 superimpositions covering the whole surface, were listed in Anati in 1960. Apart from descriptive sections, it is evident that a figure record, after being completed, appears as a series of fields filled with alphanumeric codes. That said, it is clear that there is a need to translate, manage, and compute this rather large amount of data, a fortiori if many figures are involved, as in the case of the Great Rock of Naquane. Dedicated software, RAD­ Rock Art Database (RAD.exe; Area 1997), has been compiled since 1994 by the author of the present paper to perform these 1088 4 i = < m - A n d r e a A r c a : D ig i t a l Pr a c t ic e s f o r t h e St u d y o f t h e G r e a t Ro c k in t h e N a q u a n e N a t i o n a l Pa r k fl [TlUT¿liiT-: ___ Ca‘.T i I I- El (9 3 r.w w-~* - ■ J u t * te U |«4 u ^ J u frj»'lELfri'» 5 >»:** I wargMjHsmamari.^ ==*:•■: II I1*!#" IT 111V 0 1 3 5 * » ij«, m ta in * fT* * jin lL r lB ^ iv L f r c n il ¡1 # 3* T - ( l H t l b c p V i - ! ■ P EF f 'N lt P f a K U ■- +tm L - ± l r , l n i L i 5* Ufc ^ H ~j 1-3: - c- m : n£ai Ltdacct»'* L « ic - fjrfir U K ? 1 - a fi* . I j m ' M t M E -3 ■ —! ju ■ r f L f K v l « .d -w U - -* r* m i W l t l IS CO ÄliLCa U r r K T - i n «IM 7 ■ r Hin EJ>L"3 { i n I#S4;rrl«;.tp -lö ;S ? :i:i + j;-c b rv - m t+ ii i j p n ta (r u g - h ) +-+1 to prim ' n i . l i rlBl! LIT i 'M i j j > A u iJ fL u ■ 1 3 F i g . 9. S c r e e n s h o t s o f t h e r a d - r o c k A r t D a t a b a s e s o f t w a r e , m a i n m e n u a n d f il t e r i n g o p t i o n s f o r p r o d u c i n g t h e c a t a l o g u e OF THE FIGURES (ABOVE) AND SOURCE-CODE SAMPLES, CLIPPER '8 7 VERSION (BELOW). tasks to store and count all these qualitative and quantitative data (Fig. 9). The software took its origin from the executable EUGA.exe, built two years before the RAD to produce the catalogue and the related statistics of the 5454 figures of the Rupe Magna in Valtellina (Area et al. 1995), the largest prehistoric engraved rock in the Alps. RAD.exe source code was originally written in xBase-Clipper language (Pearson 1997)5 under a DOS environment and compiled with Clipper Summer ’87, following in some way the ‘original xBase spirit’, intended to be ‘powerful, intuitive and easy to use’, as reported by the founder and developer of HMG (see below). Clipper was simultaneously a programming language — a superset of Ashton-Tate dBase m + and xBase products — and a compiler. While as a compiler it definitively has stopped evolving since 1997, as a programming language it is still extant. RAD. exe consists of 41 program files, each written for a specific function such as record editing, database browsing or indexing, and output text-files producing. Its core procedure consists of 5 Briefly detailing some notes on archaeo-computing, the old but indispensable ‘Guide to Clipper’ (NG.EXE) in Norton Guides by Peter Norton (1987), ‘a wonderful on-line reference tool (...) to cover all commands and functions through Clipper’s Summer ’87 release’ (it was simply activated on a DOS screen by keying Shift+Fl), describes and gives the syntax of 141 functions (e.g. RECNOf) to return the current record number or TRIM() to remove trailing spaces from a string), 121 commands (e.g. DO WHILE to execute loop while condition is true (.T.) or GO TO to move to a specific record), and 22 operators (<, =, >, .NOT.); these are all the ‘bricks’ of which the Clipper Summer ’87 language was composed. 3458 rows of code, whose algorithms are mainly structured in IF-ELSEIF-ENDIF or DO CASE-ENDCASE chains and SET FILTER, USE, and SET ALTERNATE TO commands, along with the definition, translation, or calculation of specific variables. RAD.exe has been compiled in 16-bit; it is still very fast when executed in the command shell of Windows operating systems, up to Windows 7. Obviously it does not run on more recent operating systems, such as Windows 8 and 10, unless utilizing DOS emulators like DOSBox, provoking thereby a considerable slowdown in its performance. In March 2015, taking advantage of the CAA conference and of the presentation of the digital work involved in the study of NAQ1, RAD-Rock Art Database was ported to 32-bit and recompiled for a Win desktop environment (Fig. 10), as well as greatly enriched in its filtering capacities. The job has been completed thanks to freeware tools, all available online, which create a 100% backwards compatibility with the Clipper Language (Esgici 2010). This mean that after installing the software, libraries, and dependencies and completing some rewriting and updating tasks such as main menu building and windows design, not a line of the old xBase-Clipper code is lost during the 32-bit porting procedure. The set of utilized tools is composed by Harbour (Harbour Project 2011), ‘a free software cross-platform compiler for the xBase superset language Clipper’, MinGW (minimalist GNU for Windows; Mingw 2012), ‘a minimalist development environment for native MS Windows applications’, and most notably by HMG (Harbour MiniGUI-Graphical User Interface; Hmgforum.Com, 2015), 1089 CAA 2015 ft WD-R«fcArD*uh)iSt(W:(i33:. FJi MiriUQI Cifekt iMr!uV. J?0(:K ¿Itt i>it itiS t e <c -i - fK-ül M 20 x n c .4 : i; i; |tr*,1.-ww-^tirt.nr; gefdftygj >->- ¡ j g t s c-rMi Zpr^OHttibtonAsraWeO Mi ■SoikAn Siiibisp i ck Wp H BAD - Car»4 0 9 *1 r^ ju ^ s (Jwro) t*4 Hc fi fKlvi* ünhibH: C tiC O G E ftE S " " " ' I.M tU T l« IU I4'P«H iM I Jj Roiftcoi«(Wi^aflnortsl C-fii.-jii *om«41 ...: * : - v - B L'füj j eaiigsey«iEjar« 9**9UT*I itf 3 ) £-.d>.:+tv* iai 3; •i-Cifflö«! Ua Ifv.j ‘lit-l+iCTVf+ifiHU: T«emg 2wetienpunl.+vr ■\ **frcpO?K!4 8B. twW*Wl ■€, HW-HM-HOfWrf * D. v n u - i ?Jw^-rr. j r V * ■F. " F I g t n gj^iTwiridi* #** n *rrm Rita* = «JftfcUS "&. iKrewft Bh. ÜViti ^ f>MMHln«W4i ' I. «pt+Jt *A>Cl-i311ri 'L !£“■+*nr.iSTrJa i^ith>-n ft*2>- Rii- £*'!QjUBiii - ■li.-Md [ Afirt;*4C|Ceg fllt f U M I ' IPftfll U t - i v i t t h f & t i Of. '"-i-'-'H¿d O J j - ■0. tttvbptf g¥kva** ■N. H t i p w * 1 ä. tig^rifrt QCHt, £{*■ W l Üttf d U M ■I rtroont^<ü£S*cji>Jt B J SM* ■ ■ 0. 11.1- ■ 2. l - U t S 4 « K d ‘iw-drwRiAN ■5, Ela' i W EPKilD ■7. 'tit. ■ ■E '.II ■P.P- P^C:h,-?2 5 QI U v f r » 1 S frtta3 2 p#tflg :{ R *C 1M S E C 12} |D, M O K f l H S 21i£G0 F i g . 10. Sc r e e n s h o t s o f t h e r a d - r o c k A r t D a t a b a s e s o f t w a r e ( c o d e t a b l e a n d f il t e r i n g o p t i o n s ) p o r t e d t o 3 2 - b it . ‘a free/Open Source xBase WIN32/GUI Development System for Windows platform’, written by Roberto Lopez. After being fully updated and tested, the RAD.exe ported to 32-bit becomes RADwin.exe, and will be created and distributed as a freeware, fully customizable for each rock art area and site. Briefly, the main functions of RAD-Rock Art Database software are related to the management of rock/figure records and to the production of instant outputs, such as the catalogues of rocks/figures and the chrono-typological listings, rankings, and stats. The management of records includes basic functions such as adding, editing, and deleting data, as well as the customization of all classification levels. With this procedure, general or specific categories and styles may be adapted to each particular situation, provided that codes and related descriptions are conveniently translated, added, or rewritten; to justify this need, it is sufficient to cite that, for instance, the European chronology is different and not suitable for a North American rock-art site, and vice versa. The main feature of RAD regarding the file output is the production of the catalogue of figures (Fig. 11). All the data from figure records are copied into a ready-to-use (and to share) html file, where codes — categories, styles, and relations — are translated into words or phrases, paragraphs are formatted, and images (a photo and a tracing/drawing for each figure) are included for the browser visualization, ready to be published and shared online or quickly reworked into a printed book. For the north part of the Great Rock of Naquane, the entire catalogue of its 1232 figures, of which 611 are meaningful, is written by RAD in less than a minute: 280 pages, 145,000 words and 970,000 types. Another instant output is produced by the creation of detailed listings and stats: a list of the specific categories (in descending order based on the presence percentage), a list of relations (associations, superimpositions, and underimpositions), and rankings of categories and styles, both general and specific, with related percentages. For the same sector of the Great Rock we note that Iron Age figures comprise 89.94% of the total, while 82.79% belong to IV2 style, the so-called pre-naturalistic style of the Iron Age from the end of the 7th to the first half of the 5th century BC. For categories, and citing only meaningful figures, we find at the top of the ranking zoomorphic (27.7%, of which 7.32% are male deer and 4.55% are dogs) and anthropomorphic figures (9.90%). One of the most powerful and effective functions is provided by filters: each output html file — the catalogue of figures and their related stats — may be fully personalized by custom filters, for example, limiting the choice to a category/style, or to some specific categories/styles, or to a sector; or excluding categories/styles and not meaningful figures; or simply typing a text-string into description fields, and all related combinations. In this way, it is easy task to produce a convenient set of specific NAQ1 catalogues of figures, for instance footprints, deer, Bronze Age or medieval figures, useful for detailed studies and comparisons. 1090 4 i = < m - A n d r e a A r c a : D ig i t a l Pr a c t ic e s f o r t h e St u d y o f t h e G r e a t Ro c k in t h e N a q u a n e N a t i o n a l Pa r k _ : ÉlBÍflíVUlL-h«! ; RAI> R O C K A R T D A T A B A SE ** . tv m C A T A L O G O F IG U R E necia X A Q O O I ccc Sigls: >‘A<2®0I-C44Î S r ttw f Q , % u i* 442 Í 2 1 2 lig u re ie C»mh flffflVlMÇ} LnclsiBiioiw 0 ". bU+îm ; nMMUiiU. SHH di tùaitn-ïiiùat butti*. C itP .fic tií A , ñ n trc p d s iú ríi., J l siile [V nsro, l a m b í a s : í W m¿Ki T k i u c í d i u s in o n e a » E u ' d i i re m » . C á le jra ria i p e c i f ì t i . 2 1 5 , A n n a Io i ríitic iío te í-ta ip * s íi:j *k r «Jn» trw lsfc a piuw tte. CtflUHtl l'.pc«í»-naPoriLfUií). firtì VU-metá V t í f . i.C. truifú tanto* r o * lu » it « w íe $ ü h ÍV ; « t f rvçw tw L ltlerie r i d e l ì z i a ç r ü n ü lo ÿ iç i ÏV S ftn a le , q u a s i tV 3 , g i m b t a n c iu J* G ra m i* ir a i E l ü a fauifta « t t m g g l i r ç , b ra e ç u . I f . .ite i m p q g w l u c j i U x i a 3 1 ^ ¡ y i o d e s t r i 1 p i w g l a c u s p id e . ^ ¿áítttM ÉttttHi 4*ü piídl AlfiHftHfi, nfrtW ft*K hÜ ir ilia * p in o ti» . StfttdfWttO 4 CM35. n il pnta&ldHAllt anche associala '.usti la ¿inule iattura * In relazione con: filile i v - E ì t ' <Stl Ferro, IV 2 }. tfi e i t e g o m 220 - A n u b lo busto piena reKangolare-Erijie-EOicaJs-, braccia alzoHe cart lancia. ed eventuale fifi. < « H + s iu e i) G W , « d U t e - i A S rs n rfc ip o m o rii tfi specifica ffVdo¡ fi e ísííL*iLurn) 3, n id if ic a H 5 ( g r u f a i di p o n iv i ìg jh o m arC d im a sparsa. di. a lile IV * E ia ' d el Ftf», ivz). di cnUforia ipfcsfici 125 - Gruppi o linn 4i puali * Seflis a. rlg ÌWft+ìWSlì A i i;aiit?0|Kspii?rifc & siile ¡V - Et«' del f«TOP IV ÎJ. # flfltpjwifl specifica 220 ■Armalo bualo pieno i¡el"£ngoiarc<tn:pe coniale. braccia ak ate cùn lancia ed estuili» Le fig. C ^ in a u n itf G i-i? , codifica S5 ( r t p i H f l i i , l u i « , ir t e trim e to tì& ta 'b di d i s d ì I V - E 1a‘ d e l Ffrrw, JVJJ, f t i!?1f gçna j p« i.ñ í$ 1í í ■Arç.* ç arapHa o macula; fia. (te n + a u tn ) C + J 7 a «KSiftca H i (gsvppi (Si p u n it l i a » o- aiaiw llm a íp arsa «3i tu l« V E ■M o ô h h , V 4) di crfeaoria specifica 22 i - G raffi e- in r ji- k recenJSi so n . pr.iir.z1i * S o p ri a: fig (wtt+PBffl)^443* :o d :fi:i B-5 lizoc'jnorfi di siile (V - Eia' d d Ferro, JV5), di cafe-fona specifica ili - C^M*ì]pidi liOCOfBpiífO; fie. (sett^emm) G44S, codifizi &¿ (zoomorfi di siile IV - Età' Je3 Ferra, di categoria specifica 252 - U t t i l t o 4 eolk» Lwi^í -í pirtdí * íta r jw íti, probifeil* c**, a w » pfit>alHle o «iroae FIG. 11. A SCREENSHOT OF THE NAQ1 CATALOGUE OF FIGURES IN HTM L FORMAT, AUTOMATICALLY PRODUCED AND FORMATTED BY RAD; FOR CATEGORIES, STYLES, RELATIONS, AND SUPERIMPOSITIONS, INPUT CODES ARE TRANSLATED BY THE SOFTWARE INTO WORDS AND PHRASES. In reply to the second question posed at the beginning of this paper, about the contribution of software tools to the archaeologist’s work, we conclude that the case of the Great Rock of Naquane study shows that a large set of electronic instruments — for graphics and data analysis — are intended as a fundamental toolbox for the careful and up-to-date rock art researcher. On the other hand, the first question, related to the challenge between man and machine, deductions, and algorithms — basically between qualitative and quantitative science — will remain unanswered, waiting for future developments and practical evidence. Acknowledgement Language corrections by Whitney Kathryn Isaacs, M.A. Bibliography Anati, E. 1960. La Grande Roche de Naquane. Paris, Masson. Anati, E. 1966. Metodi di analisi e di archivio dell’arte rupestre. Bollettino del Centro Camuno di Studi Preistorici E: 133­ 55. Anati, E. 1974. Metodi di rilevamento e di analisi dell’arte rupestre. Studi Camuni VII. Capo di Ponte, Edizioni del Centro. Anati, E. 1975. Evoluzione e stile nell’arte rupestre camuña. Archivi 6. Capo di Ponte: Edizioni del Centro. Arca, A. 1997. RAD -Rock Art Database. [Online] http://www. rupestre.net/orme/database. [Accessed: 26 June 2015], Arca, A. 1999. Digital auto-tracing in Rock Art Recording. Applications of computer vectorial design. TRACCE Online Rock Art Bulletin 11 (February 1999). Available from: http://www.rupestre.net/tracce/?p=1989. [Accessed: 26 June 2015], Area, A. 2000. Computer management of alphanumeric and visual data in the alpine rock art, Valcamonica, Valtellina, western Alps. Arkeos, perspectivas em dialogo 7: 55-74. Area, A. 2005. Archeologia rupestre in Valcamonica: Dos Ciii, un caso di studio. Rivista di scienze preistoriche LV: 323­ 84. Available from: https://www.academia.edu/7784103. [Accessed: 26 June 2015], Arca, A. 2009. Val Susa, rocce coppellate di Chiomonte-La Maddalena, Catalogo delle Schede da Sus-Madl a SusMad5 e relazione consuntiva. In A. Arca (ed.), La Spada 1091 4 i = < ] £ CAA 2015 sulla Roccia. Danze e duelli tra arte rupestre e tradizioni popolari della Valsusa, Valcenischia e delle valli del Moncenisio: 239-54. Torino, Gruppo Ricerche Cultura Montana. Arca, A., Casini, S., De Marinis, R. C., Fossati A. 2008. Arte rupestre, metodi di documentazione: storia, problematiche e nuove prospettive. Rivista di scienze preistoriche LVIII: 351-84. Available from: https://www.academia. edu/7784083/. [Accessed: 26 June 2015], Arca, A., Daudry, D., Fossati, A. E., Morello, F., Raiteri, L. 2014. Il riparo inciso di Montjovet-Chenal (AO), seimila anni e più di iconica rupestre. In R. C. De Marinis (ed.), Le manifestazioni del sacro e nella regione alpina e nella pianura padana, studi in memoria di Angelo Rampinelli Rota, Atti del Convegno, Brescia, Palazzo Broletto, 23-24 maggio 2014: 27-66. Brescia, Euroteam. Available from: https://www.academia.edu/9449764/. [Accessed: 26 June 2015], Arca, A., Fossati, A. 2006. Rupestrian Archaeology: a methodological approach to the rock engravings of Valcamonica. In L. Oosterbeek (ed.), Europreart II. Prehistoric Art Research and Management in Europe: Case Studies: 51-8. Bari, Edipuglia. Arca, A., Fossati, A., Marchi, E., Tognoni, E. 1995. Rupe Magna: la roccia incisa più grande delle Alpi (2 vol.). Quaderni del Parco delle Incisioni Rupestri di Grosio 1. Sondrio, Consorzio per il Parco delle Incisioni Rupestri di Grosio. Battaglia, R. 1933. Capodiponte: nuove ricerche sulle rocce incise della Valcamonica. Notizie degli scavi di Antichità comunicate alla R. Accademia dei Lincei per ordine di 5. E. il Ministro dell'Educazione Nazionale e pubblicate d ’accordo col R. Istituto di Archeologia e Storia dell’Arte 6, IX, fase. 7°, 8°e 9°: 201-39. Bednarik, R. G. 2001. Petroglyphs in Italian Alps dated. Acta Archaeologica 72, 2: 109-14. De Marinis, R. C. 1988.1 Camuni. In A. M. Chieco Buanchi et al. (eds.), Italia omnium terrarum alumna: la civiltà dei veneti, reti, liguri, celti, piceni, umbri, latini, campani e iapigi: 131-54. Milano, Libri Scheiwiller. Esgici, B. 2010. Vivaclipper! Notes o f a Clipper Language Student. [Online] https://vivaclipper.wordpress.com. [Accessed 26 June 2015], Fossati, A. 1991 L’età del Ferro nelle incisioni rupestri della Valcamonica. In La Guardia, R (ed.), Immagini una aristocrazia dell’età del Ferro nell’arte rupestre camuna: contributi in occasione della mostra, Castello Sforzesco, aprile 1991-marzo 1992: 11-71. Milano, Comune di Milano, Settore Cultura e Spettacolo, Raccolte Archeologiche e Numismatiche. Fossati, A. and Arca, A. 2001 Tracing the past. Petroglyph reproduction for Rupestrian Archaeology. In R. La Guardia (ed.), Secondo convegno intemazionale di archeologia rupestre. Archeologia e arte rupestre: l ’Europa-le Alpi-la Valcamonica. Atti del convegno di studi, 2-5 ottobre 1997, Darfo Boario Terme\ 267-70. Milano. Fossati, A., Jaffe, L., Sìmoes De Abreu, M. 1990. Rupestrian Archaeology. Techniques and Terminology, a Methodological Approach: Petroglyphs. Ricerche Archeologiche 1. Cerveno, Cooperativa Archeologica Le Orme dell’Uomo. Harbour Project 2011. A 32/64-bit xBase compiler. [Online] https://harbour.github.io. [Accessed: 26 June 2015], Hmgforum.Com 2015. Exclusive forum fo r HMG, a Free / Open Source xBase WIN32/64 Bits / GUI Development System. [Online] http://www.hmgforum.com/index.php. [Accessed: 26 June 2015], Loendorf, L. 2001. Rock Art Recording. In D. S. Whitley (ed.), Handbook o f Rock Art Research: 55-79. Walnut Creek, Rowman & Littlefield Publishers. Marretta, A. 2014. Tecniche di incisione e metodi di documentazione dell’arte rupestre in area centro-alpina: una panoramica aggiornata Notiziario dell’istituto Archeologico Valtellinese 12: 7-20. Available from: https:// www.academia.edu/9573735. [Accessed: 26 June 2015], Marretta, A., Martinotti, A., Colella, M. 2013. Un’esperienza di procedura documentativa e analitica informatizzata di tecniche e sequenze istoriative su due frammenti litici con graffiti protostorici da Piancogno (Valcamonica, BS). Archeologia Postmedievale 17: 319-28. Marro, G. 1932. Il grandioso monumento paietnologico di Valcamonica. Atti della Reale Accademia delle Scienze di Torino LXVE, disp. la e 2a: 413-89. Mingw 2012. Minimalist GNU fo r Windows. [Online] http:// www.mingw.oig. [Accessed: 26 June 2015], Pace, D. 1972. Petroglifi di Grosio. Tellina Opuscolo 2. Milano, Scuole grafiche pavoniane-Istituti Artigianelli. Pearson, D. 1997. The Guide To Clipper Summer 87 - Norton Guide. [Online] http://www.ousob.com/ng/sum87/index. php. [Accessed: 26 June 2015], Scott, U. 2015. Safeguarding Prehistoric Rock Art through Modeling. [Online] http://blog.pix4d.com/ post/122424602526/safeguarding-prehistoric-rock-artthrough-modeling. [Accessed 26 June 2015], Seidl M., Wieser E., Alexander, C. 2015 Automated classification ofpetroglyphs. DAACH- DigitalApplications in Archaeology and Cultural Heritage. In Press, Corrected Proof Available online, 14 March 2015. [Online] http:// ment.org/DAACH 15/DAACH-D-14-00016Rl_preprint. pdf. [Accessed: 26th June 2015], 1092 4 i = < m -