I. Delimitation of the project: periods and kinds of sources

From its discovery in 1500 to 1822, Brazil was not an independent country: it was a Portuguese colony. During this period, it is impossible to draw any clear-cut distinction between Brazilian and Portuguese science. After the independence of Brazil in 1822, there became an increasing cultural divergence between the two countries, but even in this period it is useful to compare their academic productions. For those reasons, this project was devised to include both Portuguese and Brazilian works.

For practical reasons, the project was first restricted to the period from the beginning of printing to 1822 (the year of the independence of Brazil). For this period, several parts of the project are now fairly advanced. They include data-bases on: published scientific, medical and technological books, theses, pamphlets and articles; secondary literature (bibliographies and historical studies) about this period; printed and manuscript maps; manuscripts of scientific and technological import; and biographical data on authors of those works.More recently (since 1994), the project was broadened to include the period from 1823 to 1900. For this period, however, only published works have hitherto been included.

II. Subjects included in the project

The following subjects were excluded: religion, law, fiction literature (romance, poetry, drama, etc.), politics. However, some religious works were included for their linguistic relevance (for instance: religious texts written in indigenous or Asiatic languages). A few laws, and fiction works directly related to science and technology (for instance: Gil Vicente's comedy "Auto dos físicos", that depicts the author's view on Portuguese Renaissance medicine), were also included.

Works about Portugal and Brazil written and published in other countries were not included. For instance: many foreign travelers have described the geography, flora and fauna of Brazil, but their works cannot be considered as part of the Brazilian scientific production.

History was partially included in the search. It was not desirable to include all works that might interest a historian working on the history of Portugal or Brazil (in that case, nothing could be excluded). The selection was made from the assumption that this data-base was to be used by historians of science, medicine and technology. For this reason, the search included historical works that presented an account of the general historical context of Portugal and Brazil during the period, but excluded specialized works of political import, diplomatic correspondence, discourses, laudatory biographies of kings and princes, etc.

All manuscripts that corresponded to scientific, medical and technical monographs were included, of course. Individual letters were included only when it was written by an important author, or when its content was clearly relevant. Institutional papers (those, for instance, of Coimbra University) were included as large units, without detailed description.

A very broad delimitation of "Portuguese and Brazilian" authors was used. The search included any works written or published in Portugal, Brazil and other Portuguese colonies, as also works written by authors that lived in those countries, but were published elsewhere. Portuguese translations of foreign works were also included.

III. Method of information search

Four important Brazilian and Portuguese bibliographies of bibliographies were used to find reference works . They list several thousand items. A selection of those items was used in the project.

There are several Brazilian and Portuguese general bibliographies (not restricted to any specific subject or period) that were used in the present project . There are also some important general bibliographies restricted to a specific time period . It was necessary to examine those works page by page, in search for relevant information.

For the search of printed books it was useful to use printed catalogues of important Portuguese and Brazilian libraries , and printed catalogues of book-sellers (some of them are auction catalogues of private libraries) . There are many works of this kind that were published in the 19th century and beginning of the 20th century. Unfortunately, there are no general printed catalogues from the two largest libraries: Biblioteca Nacional de Lisboa, and Biblioteca Nacional do Rio de Janeiro.

Thematic bibliographies and historical studies provided important information. There are a few good works on nautical studies , medicine , history and geography , mathematical sciences , ethnology and indian linguistics , on works published in specific languages or places , etc. Of course, they are far from complete, but provide a lot of useful information.

Some highly specific studies (sometimes on a single author) were also consulted . They usually provide fairly complete information on a very limited topic.

A few international catalogues and bibliographies were also searched for relevant information. In general, the probability of finding Portuguese or Brazilian works in international catalogues is, of course, small. However, it was useful, for instance, to consult Sommervogel's bibliography  because up to 1770 teaching in Portugal and its colonies was directed by Jesuit priests.

Many of the above mentioned secondary works contain information about maps and manuscripts, besides printed works. However, for maps and manuscripts other specific works were used.
The older Portuguese and Brazilian cartography has deserved several good studies . Catalogues of map collections , expositions  and specific studies  were also used.

In the case of manuscripts, printed catalogues of several important archives are available  and were used. There are also some useful thematic manuscript catalogues and studies .

A few computerized data-bases were also used. There is a Portuguese national net of libraries database, directed by the Biblioteca Nacional de Lisboa. Unfortunately, however, only a small proportion of old books have been entered in this database. The CD-ROM versions of the British Library and of the Bodleian Library catalogues have been consulted, as they provide the possibility of search by place and year of publication. Several important on-line databases (such as the Library of Congress catalogue) could not be efficiently used, since they do not allow primary search by place of publication or language.

The search for information on periodical articles used a twofold technique: direct search of important periodicals, and use of secondary literature. In the case of known scientific, medical and technical periodicals, the best approach is, of course, to consult the periodicals themselves and list their contents. However, in the 18th and 19th century many relevant articles were published in general periodicals - and even in newspapers. It is impossible to check all general periodicals in search for those articles. Besides that, several Portuguese and Brazilian authors published articles in foreign periodicals - and it is very difficult to trace them. Fortunately, many articles of those kinds are referred in bio-bibliographical dictionaries . There is, of course, another practical problem: Sometimes it is difficult to find complete collections of some old periodicals. Secondary literature also helps to fill those gaps.

About 300 secondary works (bibliographies, historical studies, catalogues) have already been used for the search of relevant information. The sources presented in the footnotes of this article represent only a sample of those effectively used.

Direct search for information in libraries and archives is the other way of finding out relevant works. This method has hitherto been sparsely employed, because it is too expensive and provides a relatively small amount of information. Direct search is useful, however, to complement the search based on published catalogues, to check bibliographical data, etc. Some very rare works have also been found by direct card catalogue search in some libraries.

IV. Structure and content of data-bases

The original CDS-ISIS software did not include facilities for accentuation of words. Portuguese uses a lot of accents and for this reason two resident programs were introduced to allow the easy digitization of accented words and their exhibition on VGA screens. It became even possible to use a few special symbols that were only used in old Portuguese language (such a tilde over the letters "u" and "q").
For practical and historical reasons, there are several independent data-bases:

Main data-bases:

• HTC - printed works of the Colonial period (up to 1822) and secondary works;
• MAN - manuscripts of the Colonial period;
• PES - biographical information on authors of the Colonial period;
• MAP - printed and manuscript maps of the Colonial period;
• IMP - printed books from 1823 to 1900;
• ART - printed articles from 1823 to 1900;
• PER - general information on periodicals, from 1823 to 1900.

Auxiliary data-bases:

• HIS - a chronological data-base on relevant facts and periods of the history of Portugal and Brazil;
• BIB - contains information about libraries and archives;
• DIC - establishes equivalence between old and current orthography of words that appear in the data-bases ;
• CDU - a subject data-base, grounded upon Universal Decimal Classification, intended to help searching the main data-bases .

In the case of printed works (HTC, IMP, PER, ART), each entry includes: (a) name of author(s), editor, etc.; (b) title of the work (and of the periodical, in the case of articles); (c) place and year of publication, printer or publishing house; (d) number of volumes, number of pages or sheets, illustrations, format, size (for books and pamphlets); (d) volume, pages and year (for articles); (e) date of the original work (if different from the date of publication); (f) number of the edition (if different from 1) and reference to entries of other editions and/or translations ; (g) subjects of the work; (h) language(s) of publication; (i) reference to biographical entry in PES data-base (if available); (j) reference to associated manuscript in MAN data-base (if available); (k) sources of information (code of secondary work and page); (l) localization of the work (libraries where it can be found and signature - shelf code - if known); (m) reference to image of title-page (if available) - only for pre-1822 books; (n) transcript of title-page (when available) - only for pre-1822 books; (o) other comments and observations.

The biographical data-base (PES) contains information about: (a) complete name, pseudonyms and variants of name of authors; (b) date of birth and death; (c) place of birth and death; (d) profession and fields of interest; (e) institutional links, places where the author lived; (f) reference to entries in other data-bases connected to this person; (g) sources of biographical information; (h) other comments and observations.

The cartographic data-base (MAP) contains information about: (a) author(s) of the map, if known; (b) title of the map, if there is any; (c) date of the map, if available; (d) subject of the map; (e) language used in the text of the map; (f) technical information about the map (printed or manuscript; kind of printing; size; use of colours; etc.); (g) reference to copies or different versions of the map; (h) sources of information used for each entry; (i) localization of the map (library or archive, and signature); (j) other comments and observations.

The manuscript data-base (MAN) contains the following descriptive elements: (a) author and/or copyist (when known); (b) title of the manuscript (when available); (c) date and place of the manuscript (when available); (d) subjects of the manuscript; (e) language(s) of the manuscript; (g) size of the manuscript, number of pages or sheets; (h) reference of relevant biographical entry in PES data-base (when available); (i) reference to corresponding printed works in HTC data-base (when available); (j) sources of information (code of secondary work, volume and page); (k) localization of the manuscript (library or archive, and signature); (l) other comments and observations.

Besides those main fields, there are other coded information in the several data-bases.

All data-bases allow searches using information in any field (except the "comments and observations" field). It is possible to enter any Boolean combination of parts of the title of the work, parts of the name of the author, date, place, subject, etc. For instance: it is possible, fast and easy to search for medicine works written in French, published in Coimbra in the 18th century. A special fast search is possible for decades and centuries (for instance: 176- will find all dates from 1760 to 1769). Words can also be truncated, in order to search for related terms. When the truncated word has 3, 4 or 5 letters (for instance: ast-, astro- or astron-), the search is as fast as with complete words. Accented and non-accented forms of the same word are equivalent, in any search.

The use of the auxiliary DIC data-base allows to find old forms of author names (such as "Affonso") and title words (such as "chymica") when the modern form is used in the search. This technique was found to be much more satisfactory than phonetic similarity programs . It required, however, the development of an auxiliary data-base and still requires periodical actualization.

V. Present state of the project

The HTC data-base contains entries for about 5,900 books printed up to 1822, 400 articles published in the same period, 1,200 re-editions and works written before 1822 but published after this date, and 1,900 secondary works (bibliographies and historical studies). This part of the project is fairly complete. Most of the future increase will correspond to secondary works and post-1822 publication of Colonial works.

Data-bases IMP and ART were begun only one year ago. They are increasing very fast, as always happens in the beginning of each work. If financial support is maintained, they will attain twice their current sizes at the end of 1996.

In its present state, all those data-bases are fully operational. They can be used in site, while remote access is not possible. Researchers interested in any specific theme can also contact the project by letter or electronic mail . Work will continue for several years, in order to complete, revise and improve current data-bases.

Acknowledgment

Many people have directly cooperated with this work, in several ways: searching for new information, improving the structure of the data-bases, introducing and checking data, etc. The main participants of the project have been: Maria Cristina F. de Toledo, Gisele Doratioto, Ermelinda Pataca, Paulo Henrique Teixeira, Lourdes Isabel de Lira, Andréa Regina Pereira, Poliana Barreiro, Lucila de Sousa Piva, Fábio de Oliveira Santos, Gisele Toledo da Fonseca, Naylson Maciel, Sony Fermino dos Santos, Elisa Lopes Alberto, André T. K. de Assis and Lilian Al-Chueyr Pereira Martins.

This project could not be undertaken without the help of many libraries and archives. We are specially indebted to the following ones: Biblioteca Nacional do Rio de Janeiro, Biblioteca Nacional de Lisboa, and Biblioteca da Universidade de Coimbra.