Improving a Growing Atlas

Server Logs

Basically, a server log file is a file in which every request to a web server is recorded in text format. The information recorded with each “hit” or request to the server differs depending on the configuration of the server. However, common data recorded includes the date and time of the request, the Internet Protocol (IP) address of the computer making the request, and the name of the requested resource. The file is built chronologically, and, if a web site experiences heavy traffic, the files can grow quite large. Most web masters employ a regular archiving method to prevent any single log file from becoming too large. A site might begin a new log file once a day, once a week, or once a month, depending on the level of traffic it experiences. The two most popular web servers, The Apache Software Foundation’s Apache HTTP Server and Microsoft’s Internet Information Server (IIS) have different default settings for how log files are managed. In Apache, the default is to build a single “access.log” file that can grow quite large if not managed with a regular archive program, while the default in IIS is to archive the log file daily, which can lead to large numbers of smaller files named by date.

The two web servers also record different information by default, and CWA web masters who are hoping to gather detailed information from their server logs must pay attention to the server settings in order to ensure that the server is set to record all the types of useful information that are desired for any subsequent reporting. For example, by default, the Apache web server does not record the “referring site” or “user agent” in the log. By default both these pieces of information are recorded in IIS logs, although a web master would be wise to confirm this early in a project’s deployment. A “referring site” is the site from which a user came to a CWA, such as a search engine or link in another web site. “User agent” is the technical term for the web browser application of the computer making the request. Knowing the referring site can help CWA administrators discover which external sites are driving substantial amounts of traffic to the CWA project site. Knowing the user agents can help inform on the computer platforms employed by the CWA audience (e.g., percentage of Windows versus Mac users) as well as provide a good indication of which browsers need to be tested for compatibility during design and construction. If this information is of interest, then web masters are advised to check the information being recorded in logs to confirm that they are getting what they require.