File Compression

The data files distributed by the Matrix Market are often quite large. Because of this we routinely maintain and transmit these files in compressed form. The compression program we use is gzip, which is the reason for the .gz suffix on our data files. (Note that gzip is unrelated to zip, an archive and compression utility common on PCs.)

Below we provide a variety of links which will allow one to obtain software to uncompress .gz files on a wide variety of platforms. We also provide pointers to information on how to configure your Web browser so that it automatically uncompresses these files on receipt:

On Unix systems this often requires nothing more than inserting the line "encoding/x-gzip; gunzip %s" in the .mailcap file in your home directory.

On PC systems (Windows 95/NT) some browsers continue to ignore Content-Encoding fields which can lead to failures when downloading compressed files (see Ginsbarg's flame). There are workable solutions, however; see Additional notes for PC users below.

Finally, in many cases we also make available multifile archives of matrix data which have been constructed using Unix tar. tar is available on all Unix systems (see man tar). Some of the PC resources linked to below will also handle Unix tar files.

Compression Resources on the Web

gzip	Source and executables for Cray, DOS, DEC5000, HP, IBM RS/6000, Sun, SGI. (DOS users, see note below.)
WinZip	Shareware for PCs which can uncompress gzipped files. Supports Windows 95. Also handles tar files.
UMich archive	Archive of compression software. Includes macgzip.
MacArchive	Archive of compression software for MacIntosh computers.
Netscape	How to configure Web browsers for helper applications.

Additional Notes for PC Users

Our recommendation for PC users is this: use the latest version of WinZip (v. 6.2) on Windows 95/NT. (Windows 3.1 and DOS will have problems with our long filenames.) Set the your browser's "x-gzip" option to the location of the WinZip executable. This will pop up a window when you download a .gz file. Double-click on the filename, or use the "View" button to see its contents. You can then extract this into an uncompressed file.

Notes for WinZip users:

WinZip 6.1 on Windows NT 4.0 has been known to have occasional problems. The latest version (6.2) seems to work fine.
For viewing the files inside WinZip, use the Windows WORDPAD.EXE , rather than the default NOTEPAD.EXE (this can be changed from the "View" menu). NOTEPAD.EXE cannot handle large files, nor does it handle the CR/LF conversion properly.

Notes for DOS gzip users:

The PC version of gzip above may not work with long filenames such as those found in the Matrix Market. For example, in Windows NT, gzip 1.2.4 (18 Aug 93) chokes on "gzip -c -d young1c_mtx.gz",reporting that the file does not exist. The problem is that applications that recognized only the 8.3 DOS names see something like young1c~.gz. As a result, it may be necessary to rename the file when it is saved to local disk.

The Matrix Market is a service of the Mathematical and Computational Sciences Division / Information Technology Laboratory / National Institute of Standards and Technology

[ Home ] [ Search ] [ Browse ] [ Resources ]

Last change in this page : December 3, 1999. [ ].