CORPUS OF BRAZILIAN MEDIA PORTUGUESE
         ====================================

                  Information file 
             Version of October 10 1994


This is a collection of news stories from Brazilian 
newspapers,magazines and television as they were 
distributed on the Internet. Most files are organized 
by the date of publication/broadcast and named ' s day month year '; 
thus s110194.txt means news from the 11th of January, 1994. 
When the figures for day are zeros, the file so named 
refers to several successive days, as in s000894.txt, 
which contains several individual files from the month of 
August 1994.

Files have been archived (with tar) and compressed  
(with gzip) so as to fit on 1.4MB diskettes:

file            tar bytes       tar gzip bytes
disk1.taz       3,333,120       1,203,004
disk2.taz       3,221,504       1,170,502


Because the files contain a number of words in email 
headings which are not words of Portuguese, only a fair 
estimate of the size of the text collection is possible. 
It is estimated that each disk contains about 400,000 to 
450,000 usable words.

Accentuation is not represented consistently. Only final 
acute accents appear more frequently as a single inverted 
comma.Spelling is also incorrect sometimes.


Disclaimer
==========

Since the source material for this collection has been 
distributed to the public on the Internet, the depositor 
cannot be held responsible for infringements to copyright 
laws. The responsibility for clearing copyright permissions 
rests with the users of this material and not with the depositor.
The depositor cannot guarantee the authenticity of the texts
either.