Configuration

The Next Step in Information Technology

Configuration
Comments and questions: Summary@Summary.Net
Summary is configured through your web browser. Go to the Summary main page, the URL given by Summary when it is run, and select Configuration. This will take you to the main configuration page.

Main Page
Registration Code - This is the place to enter your registration code. Your registration code will be e-mailed to you a few days after you register. See the section on Registering for more information. A valid registration code will remove the "Unregistered" message from the page heading and allow you to access the automatic processing every N hours option.
Port to serve on - This is the TCP/IP port that Summary uses to serve configuration and report requests. A value of zero causes Summary to automatically select a free port from the list 80, 8000, 7000, and 8001 to 8010. The normal port for web servers is 80. The WebSTAR Proxy service defaults to port 8000. Other port numbers between 1 and 1023 are normally reserved for existing services. 8000, and 7000 are common alternative ports. If you enter a port number and that port is not available, Summary will not be able to operate properly.
Name required to access configuration
Password required to access configuration - If you fill in either of these fields, anyone accessing the Summary configuration screens will be asked for a user name and password. They will need to enter the values you enter here to view or change configuration information. If one of the fields is blank, Summary will accept any value for that field. This user name and password can also be used to access reports. If you forget your name or password, you can quit Summary and delete the lines in the config.txt file in the config folder that start with AuthName and AuthPass, then re-run Summary and enter new values.
Site descriptive name - This is the name that appears at the top of all of the reports, normally used to describe the sites the report is for.
Preferred domain name - This should be the full domain name of the server that produced the logs you are analyzing. This name is used for making links to the pages on the server from inside the reports and also as one of the domain names considered local (see below). It should be entered all lowercase. For example "www.summary.net".
Name required to access reports
Password required to access reports - If you fill in either of these fields, anyone accessing Summary will be asked for a user name and password. They will need to enter the values you enter here to view any information from Summary. If you have also entered a configuration name or password they will not be able to do any configuration. If you have not entered a configuration password they will need this name and password to access configuration as well as reports. If one of the fields is blank, Summary will accept any value for that field. If you forget your name or password, you can quit Summary and delete the lines in the config.txt file in the config folder that start with ReportName and ReportPass, then re-run Summary and enter new values.
Other local domain names - These name are used, along with the "Preferred domain name" above, to determine if a referrer is local or remote. You should enter all of the domain names that can be used to access your server, including numeric IP addresses, except for the one you entered as the "Preferred domain name". These names should be entered all lowercase. For example;
summary.net
204.107.211.172

Filtering
Filtering can be used for several purposes. You can exclude hits you make yourself, to get a better idea of what your visitors are doing. You can filter down to only a single page, this causes the time reports to tell you about aceess to that page over time. You can filter out specific robots, such as a checker that makes sure your site is up and running, that might distort the results. You can filter on server name to get reports on a single domain from a log with several domains mixed together. All of the filtering fields take a list of items to ignore. Any log entry which has a field that matches one of the entries in the corresponding lists is filtered out, i.e not used for processing at all.
Items are compared to each entry in the list from top to bottom. The first match is the one that is used. You can use a '*' in an entry to match any string. Line which start with a '+' indicate items which should be included, even if subsequent lines exclude them. For example you could give:
+summary*
*
to include only those entries that start with "summary".
Servers to ignore - The server field is taken from the WebSTAR CS(HOST) token or Microsoft Extended logs. It normally indicates the domain name or IP address of the server requested by the user. For example "summary.net".
Hosts to ignore - The host field is the domain name or IP address of the clients. For example "proxy37.aol.com".
Requests/files to ignore - The request is the name requested by the user/browser. These are written in Unix style, ie. "/summary/download.html".
Visit initiating referrers to ignore - This is the referrer from the request that initiated the visit. This field will be the same for all subsequent requests during a single visit, regardless of their individual referrers. Normally this is the URL of an external site that has a link to your site, which was followed to start the visit. For example "http://www.infoseek.com/Topic/Macintosh_web_servers".
Auth. users to ignore - This is the name entered by the user in response to an authorization dialog. This field is typically blank unless this is a request to a restricted access portion of your site.
Cookies to ignore - The cookie string returned by the web browser along with the request. This is taken from the WebSTAR log token CS(COOKIE). This field is typically blank unless you have put cookies in your site.
Agents to ignore - The agent is the long identifying string provided by the web browser to indicate which version of which browser is making the request on behalf of the client. These strings vary quit a bit even within a single browser. For example "Mozilla/4.04 (Macintosh; I; PPC)".

Time Units
Number of hours between processing runs - You must be registered and have entered your registration code for this field to appear. Summary will process the log files automatically every this many hours if you leave it running. Set this to zero if you only want manual log file processing runs.
Number of days to include in report - Summary will only process requests back this many days from the most recent request in all of the log files. All earlier requests are ignored.
Number of days to count as 'current' - Several reports have a Curr Hits column. This setting controls how many days before the last request in all of the log files counts as being in the current period.
Number of idle minutes to end visit - A visit consists of a sequence of requests from the same host with a gap of no more than this many minutes between requests.
Number of days to keep daily statistics on - This controls how many days before the most recent request in all of the log files appear in the Daily Report.
Number of hours to keep hourly statistics on - This controls how many hours before the most recent request in all of the log files appear in the Hourly Report.
Number of minutes to offset times in logs - This number of minutes are added to every date of every request. If your logs are kept in GMT and you want a report for the east coast of the US during daylight saving time (which is four hours before GMT) you would set this to -240.

Options
Lowercase request names - Some servers, particularly on the Macintosh and Windows, ignore the case of the requests. Unix servers are normally case sensitive. Check this check box if you want Summary to ignore the case of requests.
Do DNS lookups - Web servers typically run more efficiently if DNS lookups are turned off. To allow Summary to report on domain names instead of just numeric IP addresses check this check box. This will drastically slow down processing while reverse DNS lookups are being done, but the requests will be cached for quicker access on subsequent runs.
Create CGI/Search argument report - The CGI argument report can take a lot of memory and is only useful to some people. Check this check box if you wish to keep records on CGI arguments passed in requests to your server.
Create cookie report - The Cookie Report can take a lot of memory. Most sites don't use cookies and some that do use a unique cookie every time. Check this check box if you want records on cookies passed in requests to your server.
Process logs automatically at startup - Summary normally processes the log files automatically each time you start the application. If you have very large log files you might not want this to happen. Check this check box to cause Summary to only processes logs when requested to do so by the user or by the "Number of hours between processing runs" configuration setting.

Request types
The extension of a request is the portion of the request after the last period character in the file name portion of the request, or blank if there is no period in the file name portion or the last character of the request is a period, or a '/' if the request ends with a '/'.
Page file name extensions - These are the extensions you wish to have treated as page requests. Page requests appear in the Pages report, the Entry and Exit Point reports, and are used for tracking paths through your site. Common page file name extensions include: "/", "html", "htm", "shtml", "sht", "txt", and "asp".
Graphics file name extensions - These are the extensions you wish to have treated as graphics requests. Graphics requests appear in the Graphics report and are ignored for the purpose of path tracking. Common graphics file name extensions include: "gif", "jpeg", "jpg", and "jpe".
Download file name extensions - These are the extensions you wish to be treated as downloads. Downloads appear in the Downloads report and are also treated like pages for the purpose of path tracking. Common download file name extensions include: "exe", "bin", "sit", "hqx", "zip", "gz", "z", "Z", "uu", "uue", and "tar".
All other extensions are referred to as Others in Summary.

Miscellaneous
Number of lines on a report page - This is the number of items that can be displayed on a single page of a report. A few reports, such as the Path report, use one fifth of this number. You must enter a number between five and one hundred. More lines on a page makes the reports take longer to load and display in the browser, but increases the amount of information that can be viewed without going to the next page.
Names that '/' defaults to - When the user makes a request ending in '/', the web server will look for a file with some special name or any of several special names and serve that file instead of doing a directory listing or getting an error. The exact set of file names depends on how you have configured your server, but commonly includes "index.html" or "default.html". Enter the same set of names your server is configured to use here. This allows requests to "/" and to "/index.html" to be counted together as requests to "/".
User log format definition - This is a string of tokens to indicate to Summary how to parse the log file. Summary will parse most log files automatically, and this field can be left blank. See the appendix on log file formats to find out what you should enter here to deal with special log file formats.