File Formats

Files are in several formats which all have their advantages and disadvantages. To save space and downloading time, they are often compressed. For compression we use gzip version 1.2.4. Compressed file names end in ".gz".

Common Formats

Here are some examples of common file formats, illustrated in some cases with a small (two page) abstract.

HTML is used for web pages, such as this page formats.html. They can refer to files in other formats, e.g. gif (see below).
TeX or LaTeX source files (tex) are useful for copying text but hard to convert to printable form unless you have the correct version of TeX or LaTeX and the appropriate style files.
Here is an example written in AMS LaTeX (4203 bytes) and here is the compressed version (2100 bytes).
TeX and LaTeX produce device independent (dvi) files. Compressed dvi files are usually small (hence fast loading) but may give trouble with fonts. The correct fonts may need to be generated using Metafont the first time they are needed. Here is an example dvi file (5348 bytes) and here is the compressed version (3127 bytes).
Postscript (ps) files avoid font problems but are larger than dvi files. If generated for a specific paper size (e.g. A4) or printer resolution (e.g. 600 dpi) there may be a problem printing them with a different paper size (e.g. American letter) or printer resolution (e.g. 300 dpi). Compression is worthwhile with postscript files.
Here is an example (67996 bytes) and the compressed version (30132 bytes).
Adobe Portable Document Format is perhaps the most portable but also (usually) the largest. You can download a free Adobe Acrobat (pdf) Reader (version 4.0 or higher is recommended). Here is an example of a pdf file (89131 bytes). For printing, the "fit to page" option is recommended.
Some postscript and pdf files are large because they encode images (perhaps scanned text).
Pages which have been scanned are sometimes available in GIF (gif) format. Here is an example (6919 bytes). In most cases the separate gif files for each page have been combined to give a single pdf file. Here is a rather large example (2336671 bytes).

Other File Formats

Other formats include plain ASCII text files (txt), which are useful for things such as numerical tables, since they can easily be used as input to programs. Text files are often compressed.

Here is a plain text file (2269 bytes), and here is the corresponding compressed file (663 bytes).

Here is a JPEG file (a picture, 10382 bytes).

Sometimes it is convenient to collect several files in one large file. On Unix systems this is usually done using a TAR (tar) file. It usually pays to compress tar files unless their contents are already compressed. Other systems have different utilities to accomplish the same purpose as tar, and some combine this with compression (e.g. PKZIP is common on DOS systems).

Here is a sample tar file (24576 bytes) and here is the corresponding compressed file (3633 bytes).

Format of Publications

In the list of publications:

[dvi] means a file (usually gzipped) in Tex .dvi format
(usually much smaller than postscript, but may give trouble with fonts).
[exe] means an executable file (usually a self-extracting archive for IBM PC or compatible computers)
[html] means a file in hypertext markup language
[ps] means a file (usually gzipped) in postscript format
(the page size is usually A4, but the margins should permit printing on American letter size paper which is a little shorter and wider than A4).
[pdf] means a file in Portable Document format.
[tar] means a gzipped Unix tape archive (tar) format file.
[txt] means a plain ASCII text file (possibly gzipped).
[gif] means a gif page image (not gzipped, since gif files are already compressed).

Viewing Compressed Postscript Files

Most of my papers are available in pdf format and/or gzipped postscript format ([ps] above). Most browsers can handle at least one of these formats. For information on how to view compressed postscript on various systems (including Windows, Macintosh, Unix), please press here.

Sending Files to me by Email

If you want to send me a file by email, it is reasonable to assume that I can decode and/or print files in any of the above formats (there may be exceptions due to the level, memory requirements, page sizes etc). Plain text [txt] or postscript [ps] is usually best. If you send any binary files (e.g. dvi, pdf, tar, or any compressed files) they need to be encoded before being sent by email. For example, uuencoded, "pgp -a" and "base64" encodings are acceptable.

Intentional Omission

There is no mention of Microsoft Word above. This is because the Word format is ill-defined and proprietary. I seldom produce Word documents and I have difficulty reading them. If you want me to read a Word document, please see the instructions here.

Return to Richard Brent's index page