saggi musicali italiani (SMI) is an evolving database for texts of Italian music theory and aesthetics. Following the model of the Thesaurus Musicarum Latinarum (TML), SMI makes its texts available in electronic form, allowing them to be downloaded, browsed, and searched. saggi musicali italiani focuses first and foremost on major treatises but also includes articles from journals and reviews when a distinct portion discusses a significant theoretical or aesthetic issue of general interest. SMI will eventually contain all printed materials from the Renaissance to the present.
At present, the entire SMI database may be viewed and searched online or downloaded to a scholar's personal computer (MS-DOS, Macintosh, or Windows) from two Internet resources. It may also be obtained on a CD-ROM (ordering information is provided from this link). When downloaded to a personal computer or read from the CD-ROM, in a matter of minutes even more sophisticated searching--including proximity and "fuzzy" searches--can be accomplished and displayed through two special programs, UltraFind 2.5.3 (for all Macintosh systems, including Classic under OSX) and Eureka! (for Windows98, NT, 2000, XP, or Vista). UltraFind is available from http://www.ultradesign.com/ultrafind/ultrafind.html for a free 30-day evaluation; registration (US $39.95) is required following the evaluation period. Eureka! is available from this link. The database can also be tailored to the individual's particular interests. Any part of SMI, text and graphics, will run separately or together. With UltraFind or Eureka!, searches of SMI can be endlessly configured: earlier editions of the same treatise can be included or excluded, treatises of specified centuries or a certain special group of texts ranging across many centuries can be marked for search, an individual treatise can be searched, and so on.
SMI is indebted to the TML, which granted kind permission to adapt not only the TML's system of handling graphics but also the Hypercard script for the SMI Canon of Datafiles (see below) and most of the following introduction.
SMI can be accessed in two ways: (1) by means of a World Wide Web (hereafter WWW) client connecting to a web server at www.chmtl.indiana.edu (with the TCP/IP number 220.127.116.11); or (2) through the File Transfer Protocol (FTP) at a SMI-FTP address (with the host name chmtl.ucs.indiana.edu and the TCP/IP number 18.104.22.168). Both systems will be described later in this Introduction. SMI has been designed for quick and easy access in a large number of ways, most of which should require little, if any, investment in new computer hardware or software.
The steps for using SMI are simple and straightforward (instructions for each step are provided below):
As an example, a subscription message might look like this:
SUBSCRIBE CHMTL-L Hugo Riemann [but please use your real name]
All subscription requests are forwarded by LISTSERV to the Project Director, who will in turn send a request for a short statement of interest in the SMI (this has become necessary to exclude those who would use the CHMTL-L for mass mailings or similar purposes: "spams," in the common Internet parlance). Once your statement has been received, the subscription is processed by the Project Director (this should normally take no more than a day), and LISTSERV will send two introductory mail messages: one providing general information on communications with LISTSERV, the other providing information specific to SMI. To cancel your subscription to SMI and sign off the CHMTL-L, send LISTSERV (email@example.com) the following command:
The first four steps are preliminary; it is in the fifth step where the SMI reveals its full potential.
The text, graphics, and index files of SMI can be viewed online or retrieved by means of a WWW client. This Internet option is configured in a wide variety of different ways at each site, and it is impossible for an introduction of this sort to attempt to describe the exact process one needs to follow to use the SMI Web. You should first inquire from your local specialists how to connect to your "Web home page." Once you know how to do this, you should have no difficulty with the following instructions.
When you connect to your Web home page, you are running a Web client that can link to other types of servers, which are identified by means of a URL (Uniform Resource Locator). The URL indicates the type of server (e.g., http or ftp), the location, and the path to the data. The URL for SMI is:
If you tell your Web client to load this URL, it will connect the "SMI home page," from which you can connect to a page for searching the SMI files (instructions for using the search engine are included on the page) or to the various indexes of files, corresponding to the indexes contained in the SMI-FTP. The SMI Web indexes, however, are linked to each file and function as bibliographic tables of contents. Within your Web client, you can then search the files of each index (or all the files in all the indexes), view the texts or graphics themselves, and of course save to your personal machine any or all of the files of interest to you.
If you prefer to transfer files directly to your personal machine without viewing them on the web and your mainframe or personal computer supports FTP, SMI files can be retrieved through the SMI-FTP by any subscriber using standard FTP commands.
To start FTP in a browser, type:
To start FTP in Unix, VMS, or CMS (or in a Telnet session originating from a personal computer), type:
A prompt (either an asterisk or a word or abbreviation [e.g., Command: or ftp>) will appear on the following line.
NOTE: if you are opening a Telnet session on a Macintosh computer running OSX, you should enter (without the quotation marks) "epsv4" at this point; the server will then respond "EPSV/EPRT off." If you do not enter this command, all following commands will fail.
To retrieve any text or graphics files, you must first move to the SMI-FTP by typing "cd SMI-FTP" (and without the quotation marks, of course) and then to the appropriate subdirectory filelist, which are arranged as various subdirectories of the SMI-FTP. To move down to the subdirectory from which you wish to retrieve files, enter, for example, "cd OTTOCENTO" ("OTTOCENTO" is the name of one of the subdirectories; if you wished to retrieve a file from the nineteenth-century filelist, you would use "OTTOCENTO." To move back up to the main SMI-FTP directory, enter "cd .." (the two dots tell the system to move up one directory); from this point, you can then move down into another subdirectory.
If you already know which file you wish to request at this point, simply enter the command at the prompt within the appropriate subdirectory; for example, in the SEICENTO subdirectory you might enter: "get ARTIMP.TEXT"
NOTE: in the SMI-FTP, the fn and ft must be separated by a period.
On the other hand, you may need to see a list of all the files contained in one of the subdirectories before you decide which ones to request. This is accomplished by typing (always within the appropriate subdirectory) "dir" at the prompt, which will result in something like this:
It is most likely that you will want to transfer many or all the SMI files. In this case, the "mget" command should be used (in FTP syntax, "mget" means that multiple "get" commands are being issued). Before issuing any mget commands, however, you should first enter the command "interactive" at the prompt, unless you are certain you wish to retrieve every file in the subdirectory. The "interactive" mode, which is normally set as the default for FTP servers, will cause the server to issue a separate confirmation query prior to sending each file, to which you simply respond by typing "y" or "n." If your system responds to the command "interactive" that this is an invalid command, you should then send the command "status," which should cause your system to display the default settings for the FTP transfer; one of these will normally (though not always) state either "Prompt: on" or "Confirm: on." If the status indicates that "prompt" (or "confirm") is off, sending the command "prompt" (or "confirm") will turn the function back on (see sample session below). To test the setting, try sending the command (for example, in the CINQUECENTO subdirectory) "mget AARLUC.*"; in this command, the asterisk functions as a wild card, indicating that the SMI-FTP should send all files with the fn AARLUC, regardless of the ft. If your system supports the mget option and "interactive" mode is on, this command will cause the SMI-FTP to send you the AARLUC text file with all its graphics, each file preceded by a prompt asking for confirmation before it is sent.
If you have determined that your system supports the mget option and you wish to transfer all the files in a given subdirectory, two wild cards should be used in the command: mget *.* This command will cause the SMI-FTP to send every file in the subdirectory. If the interactive mode is on, each file will be preceded by a confirmation prompt; otherwise, the files will be sent one after another with no further prompting.
NOTE: Most subscribers will find that their FTP systems are set to transfer data as ASCII text. In a few cases (especially at European sites), the data may arrive in some unrecognizable form because the default for the site has been set to something other than ASCII. If in doubt, at the FTP prompt, type "status" and press return; FTP will then display its current settings (if this does not work, typing "type" should cause FTP to respond with "a" [for "ascii"] or "i" [for "binary"]). If the transfer type is something other than ASCII, it can be easily changed. At the FTP prompt (and prior to transferring the files), type "ascii" (some users may need to type "type ascii") and press return. This will set FTP's file transfer type to the form needed for SMI files. You can check to be sure the change has been made by once again typing "status" or "type" (and pressing return) at the next FTP prompt.
SMI Text Files:
Unlike other types of texts commonly studied by scholars in fields such as classics, literature, and philosophy, those on music theory include abundant figures and musical notation for which no ASCII equivalents exist. This material cannot simply be omitted. Musical notation included within sentences is entered as codes in the text file, while full musical examples or figures are scanned and saved in GIF format and keyed to locations within the text files themselves. If the example includes text, this is given in the ASCII file within brackets (e.g., [Asioli, Il maestro di composizione, 2-3; text: CLAVICEMBALO, CONTRABASSO TEDESCO, CONTRABASSO FRANCESE, CONTRABASSO ITALIANO, FAGOTTO, VIOLONCELLO, TROMBONE, BASSO CANTANTE, TIMPANI, BARITONO CANTANTE, Suoni gravi, Suoni bassi, Suoni medj, accordatura in quarta, idem: accordatura in quinta, Voci di petto, Voci centrali e migliori, ECCETERA]), thereby enabling the search engine to locate and display text strings that appear within figures as well as those within the treatise proper. The text, of course, will also appear in the graphics file that will store the figure, table, or musical example itself.
Simple musical examples, that is, monophonic excerpts or illustrations without essential explanatory symbols, are encoded according to Barry S. Brook's "The Simplified 'Plaine and Easie Code System' for Notating Music: A Proposal for International Adoption," Fontes Artis Musicae 12 (1965): 156-60. (A description of this system is also available in Barry, S. Brook, "The Plaine and Easie Code," in Musicology and the Computer. Musicology 1966-2000: A Practical Progam [Three Symposia], edited by Barry S. Brook [New York: The City University of New York Press, 1970], 53-56.) This useful system of encoding music was designed mainly for indexing purposes and thus pays very little attention to the issue of spaces that may separate the various parameters. In our context of electronic searches, however, spacing is a crucial matter. A search for '4CDEC/CDEC/EF2G (the incipit of "Frère Jacques") would be successful only if the melody had been encoded without any spaces and not as, for example, '4CDEC / CDEC / EF2G. We thus clarify Brook's system to the effect that any spaces and all optional codes will be omitted. If a monophonic example includes text, as is usually the case in vocal excerpts, the text is entered following the codes, separated by a semicolon and a space. Codes and text together appear in brackets and are always preceded by a carriage return (ASCII 13). If a source, for example, quoted the first four measures of the vocal line of the Duke's "La donna è mobile" from Verdi's Rigoletto, the musical example would be encoded as follows (please consult Brook's article for full explanation of the codes):
[(#FCGDA,3/8)'8#D#D#D/6.#F3E4#C/8#C#C#C/6.E3#D4B/; La donna e` mobile qual piuma al vento, con brio]
Bracketed information following the codes will provide the fn and ft for the graphics file (see below), which in examples with text allows the user to examine the text underlay.
Brook's code for cut time (a "C" with a vertical line) does not figure among the basic ASCII character set. This and all other symbols characteristic of mensural notation (relevant in Renaissance and Baroque treatises) will be borrowed from the TML's "Table of Codes of Noteshapes, Rests, Ligatures...," a system of encoding developed by Thomas J. Mathiesen and published in "Transmitting Text and Graphics in Online Databases: The Thesaurus Musicarum Latinarum Model," Computing in Musicology 9 (1993-94): 33-48.
Orthography follows exactly that of the source even if a word is obviously misspelled. Three exceptions apply: (1) abbreviations are expanded; (2) apostrophes are always followed by a space as, for instance, in "l' amore" or "quest' aura," unless they replace the elided vowel of the second word, as in "se 'l creda"; and (3) all Italian acute accents are normalized as graves. Italian authors and printers handled this accent inconsistently, and normalization facilitates the process of electronically searching the texts. Of the diacriticals in Italian, only the grave accent figures among the basic ASCII character set (ASCII 32-126). The files distributed by File Transfer Protocol (FTP) thus respect only the grave accent, displaying it after the vowel to which it pertains (for example "ahime`" and "fedelta`"). All other diacriticals are ignored. On the World Wide Web, however, HTML codes make it possible to display all vowels in the appropriate form.
The graphics files, by their nature, are somewhat more complex. The GIF format has been selected in preference to any other format for several reasons. First, GIF files are quite small, and thus they can be downloaded very quickly, even over a relatively slow dial-up connection. Second, the format can be read on any of the major hardware configurations with simple conversion programs available as free- or shareware. Third, the graphics files can be displayed directly online by WWW clients.
In an effort to make the use of graphics files as convenient as possible, the TML-FTP offers two necessary programs for Macintosh subscribers running systems prior to OSX: "GIFConverter" and "UUTool." Both are stored in binhex encoding to insure that they can be transferred to any remote system without corruption. A number of GIF viewers and UU decoders are also available for Windows. There is no single common program for all these machines, but we will assist any subscriber in recommending viewers and in providing a UU decoder on disk. Please contact the CHMTL Project Office for information.
Macintosh subscribers (running systems prior to OSX) interested in viewing graphics should first retrieve the following files from the TML-FTP (the list and site of the TML): GIFCON HQX, GIFCODOC HQX, and UUMAC HQX. GIFCON HQX is a binhex-encoded version of the program "GIFConverter," an excellent shareware program by Kevin Mitchell that enables GIF files to be viewed as well as translated into formats that can be used in word processing programs or other contexts; GIFCODOC HQX is the binhex-encoded user's manual for the program; and UUMAC HQX is a binhex-encoded version of the program "UUTool," a simple but essential program that will enable the graphics files to be decoded so they can be viewed with "GIFConverter."
NOTE: SMI provides neither GIFConverter nor UUTool. These two programs must instead be downloaded from the TML in the same way as files are retrieved from the SMI-FTP, that is, by sending the appropriate request (e.g. for the GIFConverter):
After these three files have been retrieved, they must be decoded with a binhex program. When decoded, GIFCON HQX and UUMAC HQX will become self-extracting archives that contain "GIFConverter" and "UUTool" (together with a short instruction manual for "UUTool"); GIFCODOC will become a Microsoft Word file. The two self-extracting archives will show little icons like filing cabinets; when you double-click on these, your computer will ask you where to save the programs and will then extract the files for you. "GIFConverter" and "UUTool" will then operate like any other Macintosh program. Macintosh subscribers who prefer to be sent these programs on disk may request them from the CHMTL Project Office.
NOTE: The TML is making "GIFConverter" available to subscribers with the permission of their authors, but the programs are shareware programs. If you decide to use either of them, you must send the authors their very reasonable fees. The programs themselves provide you with full information.
Transferring graphics from the SMI-FTP is somewhat more complex. The graphics files on the SMI-FTP are stored in UU encoded format to insure that they can be transmitted by FTP without being corrupted. UU encoding has been used (rather than binhexing) because it offers greater standardization and can be readily decoded on any hardware platform, including Unix mainframes.
The procedure for retrieving and viewing a graphics file from the SMI-FTP is as follows. After the file has been received to a personal computer, it must first be decoded with "UUTool" or a comparable UU decoder. Once the graphics file has been decoded, it can be read by a GIF viewer such as "GIFConverter" (be sure to select the option "Look inside all files" in the "Open ..." dialogue box; then click on the button marked "fix file types for shown files"), and the graphic material will be displayed on the computer's screen. However viewed or retrieved, the graphics files are in each case coordinated with specific treatises. Thus, the file's fn matches that of the corresponding treatise, and the ft is 01GF, 02GF, 03GF, etc.
SMI will continue to adapt its treatment of the graphics portion of the database to maintain the broadest and easiest access. Users with graphics capabilities on their machines will be able to search and retrieve text as well as musical notation and figures, while those without graphics capabilities will still always be able to search the text database.
The SMI Canon:
The SMI Canon is available as a HyperCard stack indexing all the treatises in SMI, which may be retrieved from the SMI Web (by clicking here) or the SMI-FTP (by retrieving the binhex-encoded SMICANON.FULL). Each entry in the Canon includes the following fields (in this order): the name of the author of the treatise, as given in the source from which the data was taken; the author's given name, if applicable; the title of the treatise; the incipit; the source of the data file; the names of the persons responsible for entering, checking, and approving the data; the filename; the filetype; the filelist; the size of the file in kilobytes; annotations (including specific details on manuscripts, if the file is not derived from a printed source); and the type of source (i.e., manuscript or print).
Please note that the HyperCard stack can only be used by Macintosh users. It must first be decoded before it can run. This will normally be done automatically by web browsers and FTP programs, but if necessary, it can also be done by using the binhex decoding options in StuffIt, CompactPro, or any other binhex decoding program.
Some parts of the SMI Canon are also available as a PDF file, which will display an on-screen print of the contents of the HyperCard stack. PDF files can be viewed on any computer with Adobe Acrobat Reader (available for free at http://www.adobe.com/products/reader/) and are also fully searchable. The PDF version of the SMI Canon may be retrieved from the SMI Web (by clicking here).
If your personal computer is connected directly to the Internet, either through a hardwire, SLIP, or PPP connection, you will most probably have retrieved SMI files directly to your individual machine and need not peruse this section, which applies only to users who make use of a modem to effect a dial-up connection to a mainframe (or some similar intervening machine).
After the data files of SMI have been transferred to your local system, they are ready to be read and searched. Although most mainframes do have some sort of searching capabilities within their text viewers or editors and these may be suitable for certain users, SMI assumes that most users will want to download the files to their personal computers in order to take advantage of the search engine described in the next section.
Every mainframe will have a different selection of error-correction protocols available, which help insure that data transferred over telephone lines through low-speed asynchronous connections is not corrupted in the process of transmission. Most mainframes should have at least Kermit and Xmodem. Each communications program and modem addresses these protocols in different ways, and this Introduction cannot substitute for your instruction manuals. There are, however, a few points to keep in mind when you prepare files for searching.
Note: the graphics, SMI Canon, and program files, which are encoded with binhex or UU encoding, must be downloaded without any filtering or paragraph conversion.
As noted above, the SMI files can be searched on the SMI Web, and while such searches may be sufficient for many purposes, even more sophisticated searching is possible when the data files of SMI have been transferred to a personal computer. Although the files may be searched or otherwise manipulated using any one of a number of programs, at present SMI recommends UltraFind as the engine to be used for searching the database when it resides on a Macintosh (running any MacOS, including Classic under OSX). UltraFind is available from http://www.ultradesign.com/ultrafind/ultrafind.html for a free 30-day evaluation; registration (US $39.95) is required following the evaluation period.
UltraFind provides both simple and Boolean structures, as well as retrieving information even in unspecified variant spellings--an important advantage when searching pre-1900 Italian texts.
UltraFind provides an online help manual with instructions for its use and configuration, and these instructions need not be repeated here. A few recommendations, however, may be helpful.
After the desired material has been located, any number of further actions are possible: a detailed report may be printed, showing the number of "finds" and their location (in as large or small a context as may be specified); passages located may be readily imported into a word processing document; the entire text of the treatise containing the passage may be opened and printed; and so on.
As the engine to be used for searching the database on a Windows machine (runnning Windows98, NT, 2000, XP, or Vista), SMI at present recommends Eureka!, a program developed by the CHMTL. Persons interested in purchasing a copy of Eureka! should retrieve the order form from the CHMTL by clicking here.
Because Eureka! was developed specifically for use with the full-text databases of the CHMTL, the brief instruction manual included with the program will provide all the necessary information on its use. Like UltraFind, Eureka! allows for complex Boolean searches, including proximity searches; the production of detailed reports showing the number of "finds" and their location (in as large or small a context as may be specified); and nested searches.
Because SMI is composed of ASCII text and GIF graphics, it can be used by any machine with its own search program. This structure also makes it possible to adapt the files to other systems of delivery that may become feasible or popular in the coming decades, such as the CHMTL CD-ROM.
Exchanging Mail with Other Subscribers to the CHMTL-L:
In addition to its use for the announcement of new files in the database, the CHMTL LISTSERV also supports the exchange of mail among subscribers through the CHMTL-L. Questions or other communications that one may wish to share with all subscribers should be sent to:
Communications sent to this address are immediately forwarded to all other subscribers to the CHMTL-L.
NOTE: Mail intended for distribution to everyone subscribing to the CHMTL-L must be addressed to:
By contrast, requests to subscribe or unsubscribe must be addressed to:
because these are commands to LISTSERV and not mail for distribution.
Contributions to SMI:
Text files of treatises are regularly contributed to the project by scholars from around the world, and SMI very much welcomes contributions from individual scholars as well as from the members of the SMI Project Committee. Contributions of this sort should be sent to the Project Director, either over the mainframe or by mail. Please do NOT send treatises or other data files for distribution as mail on the list.
If you have any problems with SMI or suggestions for its improvement, please contact the Project Director, Andreas Giger (telephone:  578-3688; INTERNET: firstname.lastname@example.org; mail: Louisiana State University, School of Music, Baton Rouge, LA 70803).
| Return to Main Page |