J. Malitz, Gnomon. Bibliographische Datenbank. Internationales Informationssystem für die Klassische Altertumswissenschaft. CD-ROM. 3rd ed. Munich: C.H. Beck, 1996. 598 DM (498 DM for subscribers to Gnomon). ISBN 3-406-41085-5.
Reviewed by John Tamm, Department of Classics, McMaster University.
The Gnomon Bibliographische Datenbank CD-ROM (hereafter GBD) is a powerful and useful bibliographic tool for those dealing with antiquity. In some ways it is still a work in progress, both in terms of the database and its access program, but this should not discourage potential users. In the access program, only some operations not essential to the use of the program remain to be implemented. The database already contains over 180,000 entries and will expand in future editions, both by continuing to incorporate the latest scholarship as well as by filling in existing gaps.
Included in the program's database are references to all the reviews ever published in Gnomon (1-65, 1925-1995 for this edition), as well as references to the "Personalnachrichten," obituaries, and for 1990 onwards, the quarterly bibliographic appendices. The program, however, is much more than just an electronic index to Gnomon. The database also contains references to the contents, reviews and articles, of a large number -- the list takes up 16 pages in the manual -- of classical and related periodicals. In some cases, eg. ClQ, Historia, JHS, JRS, ZPE to name but a few, the complete run of the periodical has been included; for others, a subset of the complete run with an emphasis on recent volumes. For certain periodicals, namely the Anzeiger für die Altertumswissenschaft, AJA, ClRev, Göttingische Gelehrte Anzeigen, JHS, and JRS, references to all the reviews published since 1987 are included. The scope of the database is not limited just to periodicals. References to the contents of anthologies, i.e., festschrifts or conference proceedings, and series, e.g., Ancient Society and History, are also included, as are references to some monographs. This part of the database, however, is currently far less comprehensive than the journals section. Finally, there are references to over 8,000 German dissertations, from 1902-1995. Each entry in the database is broken down into a number of data-fields, filled where relevant: author (or editor), co-author (or co-editor), year, title (includes subtitle if present), textfield (keyword(s) taken from titles), place of publication, publisher, periodical, anthology title, series title, volume number, issue number, pagination and format, dissertation index number, Gnomon bibliographical supplement reference, type of publication, and descriptors (GBD-defined keywords).
The access program lets the user search this mass of data in a variety of ways (see below for more details). The result can then be displayed on the screen in one of three user-selectable formats. It is possible to sort a result in either a GBD-defined or user-defined manner. A result can also be printed out, in whole or in part. Finally, a result can be exported to the hard drive or a floppy.
Those readers familiar with DYABOLA may wonder about any overlap between the contents of the respective databases. References to all archaeological reviews published since 1956 should be accessible through DYABOLA, but apart from this, the contents of the databases are different enough that the two programs are not interchangeable. DYABOLA is an archaeological database; the GBD does have some archaeological material, but its strengths are in the fields of philology and history, areas which for the most part lie outside DYABOLA's scope. Furthermore, the GBD includes pre-1956 material, whereas DYABOLA does not. And, the GBD has at present placed special emphasis on the contents of periodicals. For comprehensive coverage of the whole bibliography, both programs are needed.
In terms of equipment, an IBM PC or compatible equipped with a CD-ROM drive is needed, running Windows 3.1 or higher. (For users without CD-ROM drives, a DOS version is available at no extra cost.) A mouse is supported, as are laser printers -- other software producers should take note. Because searches are run off the CD, a fast drive will be advantageous. At McMaster the program is installed on a Dell Dimension XPS P90 90MHz Pentium-based system, with 16MB RAM. Even with only a 2x CD-ROM drive, search speeds are acceptable. A user will also find some familiarity with German to be an asset. The language of the program is German -- menu options, dialogue boxes, program messages are all in German, as is the online help file. The manual, however, is bilingual, English and German, and does translate the terms used in the program. Nevertheless, the makers might wish to consider a multi-lingual version, to better serve their foreign customers.
Installation: Installation was very quick and easy, taking no more than half a minute on our system. Only the program and help files are copied to the hard drive; the data remains on the CD, from where it will be accessed. The installation program by default creates a new directory (c:\lidoswin) for the GBD, but the user may change these parameters. If a mistake is made, e.g. entering an invalid directory name, the user is asked to correct it. In general, the installation procedure and screens will be familiar to all who have loaded programs in Windows 3.1.
Manual & Help files: The manual is generally clear and very helpful, although not every last detail is discussed. The text is augmented by pictures of the relevant screens, and covers the main program functions and features. A detailed online help file supplements the manual; between the two the user is well looked after. There is also a series of predefined "sample tasks" which the user can load and execute. These are related to the manual, and illustrate the types of searches possible.
Look & Layout: The GBD has a standard Windows look to it, with menu and button bars across the top of the screen, scroll bar down the right-hand side, and a status bar across the bottom. Many of the program's operations can be activated by either the keyboard or mouse. The makers might wish to consider adding floating help boxes, such as appear in many other Windows programs when the mouse pointer is held over a button and which give the function of that button. Currently, if one clicks and holds on a button, an explanation does appear in the status bar. This works, but if a user doesn't remember to move the mouse off the button before releasing, s/he will end up activating a possibly unwanted process. Similarly, the cursor does not change to the usual Windows hourglass when the program is working at a task. The term "Suche" ("search") does appear in the status bar if the task is a search, but in general, the hourglass would be clearer and more familiar.
Searches: In general, searching is fast and easy. The user can have up to three search results available simultaneously, as the program provides three separate files (named A, B, and C) into which a search results are stored. "A" is the default; the user can choose the file to be used. There are three main ways to search the database. A "search in indexed fields" allows searches by author, periodical, anthology, series, or textfield. For an author search, the user can enter a single name, or a range. The search can be limited in time, to a single year or a range of years, if desired. The result will list all the works by the author(s) desired, all the reviews of his/her works, and relevant "Personalnachrichten", to the extent that these have been included in the database. For a textfield search, the user enters a keyword to be searched for. As there is no list of possible keywords, the user is left to his/her own devices. Some experimentation is needed to discover the potential of this option. For the remaining searches, the user enters the appropriate name. In each case, the result lists the contents of the desired publication. For many periodicals, however, a "descriptor search" would be quicker.
A "descriptor search" allows searches based on a descriptor or combination of descriptors chosen from a list (termed the "thesaurus" by the GBD). The user can choose to have the list displayed either in alphabetical order, or arranged hierarchically. The latter arrangement begins with broad categories and then proceeds to more specific terms. It is also possible to combine several descriptors into one search request, using the logical operators AND, OR, and NOT. The user need not use the list; a descriptor (or unambiguous part thereof) can be directly entered into the search request if desired, and this is in fact the default setting. There are currently over 4000 descriptors; the main focus is on the fields of history and philology. This is the most efficient search method, once one has learnt the system. The makers are particularly proud of this search method, and justifiably so. But as they admit, it does take a while to learn. Fortunately, some help is provided. The CD-ROM has two files which contain all the descriptors, arranged alphabetically in one and hierarchically in the other. These can be displayed and/or printed out (at roughly 135 and 85 pages respectively if one does not modify them) through any word processing program, allowing the user to become acquainted with the descriptors.
A few modifications could, I think, be made here. Currently, if the user is choosing the descriptor(s) from a list, there is no opportunity for the user to simultaneously view the search request as it is being built. In a one-term request this might not be so important, but if the request is complex, the chance to keep track of it would be helpful. In the direct entry method, the request is displayed on the screen. In both cases, an "undo" feature would be helpful; as it stands, one has to cancel the search request altogether and start a new search if one notices a mistake. Users will make mistakes, especially when they are new to the program, and they should have the chance to correct these as easily as possible. Finally, a "free text search" allows searches in which the user enters the term(s) to be searched for and the field(s) to be searched. This option is meant to cover those instances in which no keyword(s) or descriptor(s) are available. The whole database is searched under this method, so it is the slowest; searches taking up to a few minutes can be expected -- a search for the string \Hoby\ in the title field took roughly 2.5 minutes.
The ability to cancel a search once started would be useful. Rebooting the machine is currently the only option, but besides being unacceptable it seems very messy. When I tried this, I found that the Gnomon icon had disappeared when I re-entered Windows, and the only course of action proved to be reloading the program. Luckily, this does not take long, but all the same, a "cancel" button would be handy.
It will eventually be possible to save search results to disk for later reloading into the GBD. At present the menu options exist, but have not been implemented.
Display, Printing, Exporting to disk: Once a search has been executed, the result is stored in the default or user-chosen file. Sometimes, the result is automatically displayed on screen, at other times, the user has to specify this step. Again there are three methods. For each entry in the result file, the "Dokumentenliste" option displays the full database entry, the "Titelliste" option gives the author, title and year, and the "Autorenliste" option gives the author, co-author (if present) and year. The user can switch at will between these options. A result file can also be sorted, in one of two ways. In a "standard sort", the entries will be sorted by author and year. In a "free sort", the user is asked to select from a list the sort fields to be used, and the direction of the sort for each field chosen. It is possible to delete unwanted result files, and unwanted entries from within files. There is no "undo" feature though, which can be annoying.
The GBD supports printing on laser, ink-jet, and dot-matrix printers; the user will be asked to choose the appropriate format. It is possible to print out whole result files, or individual entries. The procedure for the latter is somewhat cumbersome at present, as the user has to in turn mark the entry and then choose the "selective printing" option from a menu for each individual entry one wants to print out. Within the print process lies the opportunity to export a result to the hard drive or a floppy, so that it can be used in a word processor. Three formats are available: ANSI, RTF, or DOS (ASCII). The user should experiment to find the format best suited to his/her word processor. Our WordPerfect 6.1 read both of the latter two formats perfectly; accented letters were not transferred properly with the first format. The user is also asked to enter the filename under which the result is to be saved (again, other software producers take note).
Conclusions: In general, the quality of the database is high. The makers acknowledge that there can be misprints, and are working to eliminate them. It might therefore seem petty to mention any I have found, but one of them is too inspired to ignore: John Boardman's 1989 JHS article on "Herakles, Peisistratos and the Unconceived" (rather than merely "Unconvinced"). It is worrisome that this is not a case of an incorrect letter, but an entirely different word. There may also be some duplication of entries. A free text search found, for the string \Hoby\, two references to Mueller's 1994 JdI article, once as C.W. and with a misprint in the title, and once as Carl Werner. This is the only such example I have found though.
Crashes were rare; one occurred in rebooting the system after trying in vain to stop a search. Another occurred when the "zuruck" menu option was chosen during display of a search result, after deletion of a result entry. Usually, this option acts as an "escape" feature; this time, the program was terminated and Windows had to be exited and re-entered before the GBD could again be run. No problems were encountered when running the standard program operations.
The potential for overlap with DYABOLA is somewhat worrisome. There is a niche for the GBD, especially if it moves aggressively to fill the gaps left in DYABOLA's database. Archaeologists are well served by the latter; a similar tool would no doubt be much appreciated by philologists and historians. Such a demarcation of territory would also allow potential purchasers to pick the program most suited to their needs, if they can not afford both. Some duplication will be inevitable, but I'm not sure that direct competition in terms of database contents would be in anyone's interests. (Concerning pre-1956 archaeological material, perhaps that should remain for DYABOLA to incorporate). In other aspects, DYABOLA could well learn from the GBD in terms of being a Windows program, support of a mouse and laser printer, and price. There is also the Annee Philologique to consider. Its strength is its coverage of all subjects, thus it overlaps both DYABOLA and the GBD. However, the software packages are more up to date and, I think, easier to use.
Judging a work in progress is difficult. The GBD can be a little rough around the edges, but much of that is cosmetic and easily curable. The incomplete state of the program could be the main culprit, so a final verdict must be postponed until the currently non-functional operations have been implemented. Similarly, we should wait to see in what directions the database will be extended. As it stands though, the program does do what it promises to the extent it is capable, quickly and well. It is also, for the most part, easy to learn and use. The price too is very reasonable, given the amount of information accessible. One looks forward with interest to future editions of the program, and especially, the database.
Gnomon Online: The makers of the GBD have also set up a page on the Internet, at http://www.gnomon.ku-eichstaett.de/. Here one can find additional information on the program, as well as an Internet version (Netscape 2.0 or higher is needed). This is a slimmed down version of the CD-ROM release, with fewer features, and limitations on some of the features that have been included. It does however use the complete descriptor list of the full program. It is not meant to replace the CD-ROM version, but it does serve as a handy reference tool and as an introduction to the full program. The database for it is updated weekly, and emphasises recent publications. The descriptor list can be downloaded.
From this page it is also possible to access a version of "Latin Inscriptions -- The Internet ReleaseÓ as well as information on the "Inscriptiones Graecae Eystettenses -- A Database for the Study of the Greek Inscriptions of Asia Minor". The former currently allows Internet access to over 50,000 inscriptions: the whole of Dessau's Inscriptiones Latinae Selectae, the Année Épigraphique for 1988-1992, and a selection from the CIL. It must be noted that the inscriptions are presented here in just the bare text, without bracketing (brackets make searching difficult) or commentary. Much caution is therefore needed in using this program, and the hardcopy publications remain indispensable. Nonetheless, the Internet version is easy to use, works relatively quickly, and has great potential. Both programs have been developed by J. Malitz and his team, and are available for IBM PC's and compatibles. I would recommend all interested readers to visit the site for fuller details.