Graphic Identity

File Formats

In the interest of making information readily available to as wide an audience as possible, material intended to be viewed online should be prepared in HTML format (for text and tables) and GIF (for graphics). JPEG format may be used instead of GIF for photographic material. In such cases, JPEG produces smaller files with minor loss of image definition.

The choice of file formats used should be based on the following considerations: (1) the intended use of the material by the target audience; (2) the accessibility of the format to the target audience; and (3) the level of effort required to convert the material to the format.

Material intended to be downloaded for offline print or display should be prepared in one of the following formats, which are listed in descending order of preference:

HTML and GIF or JPEG — Same as material for online viewing.

Adobe Acrobat (.PDF)— Include a link to a downloadable free viewer whenever possible.

Microsoft Rich Text Format (.RTF) — RTF is easily created from most word processors and is more widely usable than native word processor formats such as Microsoft Word or WordPerfect. However, its reproduction of fonts and page layout can vary depending on the user's font set.

Proprietary Formats (e.g., WordPerfect, Microsoft Word, Excel, PowerPoint, Freelance, etc.) should only be used if: (1) conversion to one of the above formats is not feasible; (2) the intended audience is known to have ready access to software which can handle the proprietary format; or (3) the intended use is data analysis or manipulation (see below). If use of a proprietary format is unavoidable, use an earlier, more widely available version if possible (e.g., WordPerfect 5.1, not 6.1).

Material intended to be downloaded by the user for offline analysis or manipulation should be prepared in one of the following formats, which are listed in descending order of preference:
Raw Data Files — comma delimited file, text in quotes; fixed format (mainframe data); or space delimited (purely numerical data).