Including a page hit count in a web page

Note: we don't recommend including page reference counts. The counts are inevitably inaccurate, and generating them requires additional work for the web server, with very little benefit. However, if you insist ...

There are several ways to get a page count in your HTML page. In all cases, the basic idea is that an access to your web page causes a script to be run which updates the current count and returns the new value to be included in your page. The most common ways to do this are:

• Use "server side includes". Darkwing and gladstone are configured to recognize special constructs inside (some) HTML files, and to process them on the fly.
• Include an IMG tag in your web page, but instead of a .GIF file, point to a CGI script that generates a GIF file containing the current count.

Cookbook #1 (server-side includes)

Scripts like these depend heavily on the details of the particular server configuration. The following easy (but not necessary good) technique is specific to darkwing.uoregon.edu or gladstone.uoregon.edu.

1. Create a subdirectory public_html/cgi-bin. In that directory, place the following simple program. Let's call it simplecount.cgi:

    #!/bin/sh
    FILE=`basename "$1"`.count
    echo "content-type: text/plain"
    echo ""
    read COUNT < $FILE || exit
    COUNT=`expr "$COUNT" + 1`
    echo "$COUNT" > $FILE
    echo "$COUNT"

Make sure that your program is executable, e.g. by giving the command:

    % chmod u+x simplecount.cgi

2. Let's call your page something.html or something.shtml. .shtml is easier, but doesn't work if you want a count in your index page, which has to be index.html. In your page, include some text such as:

    You are visitor number <b>
    <!--#include virtual="/~yourid/cgi-bin/simplecount.cgi?mycount" --> </b>.

Replace yourid with your darkwing/gladstone username, simplecount.cgi with the name you gave your program, and mycount with a name for this counter.

3. Create a count file corresponding to your counter to hold the current page count. If you are calling this counter "mycount," then in your cgi-bin directory:

    % echo 0 > mycount.count

Note that on darkwing/gladstone the script runs as you, and so has full access to all your files (a security problem if you have errors in your script). On other systems a similar script might run as an anonymous user, so you would need to create a world-writeable count file -- a security problem since then a malicious user could arbitrarily change the file.

4. Normally, only .shtml files get the special (and expensive) processing needed to recognize server-side includes such as the above. If your file is something.html, create a file name .htaccess (note that since the name starts with a period it will not normally be shown by an ls command) in the same directory as your page, and put in it the single line:

    AddType text/x-server-parsed-html .html

As an example, this page has been visited 11490 times as of Friday, 10-Oct-2008 23:23:15 PDT.

Cookbook #2 (cgi scripts in an IMG tag)

Writing a program that you can use in an IMG tag, and that returns a graphical counter, is not as trivial as the above textual program. However, there are various programs available for you to copy. Even easier, there are service providers on the Internet who will maintain your counts for you (e.g., http://www.digits.com/web_counter/). After initializing the counter, you need only make sure that your web page has an IMG tag that refers to the appropriate counter service. In many cases they provide not only a count, but also additional information for the page owner. For example, the page owner might be able to query the distribution of hits by type of web browsers.

An example of a very simple counter service is one provided by JQ Johnson. It is available only for pages at UO. You can use it by placing the following 2 lines on your web page:

    <IMG ALT="many" width=56 height=16
     SRC="http://darkwing.uoregon.edu/~jqj/cgi-bin/counter.cgi?
http://darkwing.uoregon.edu/~jqj/publishing/pagecounts.shtml">

Replace http://darkwing.uoregon.edu/~jqj/publishing/pagecounts.shtml with the URL for your page.

For example, this counter shows that this page has been visited times

Caveats (some problems with doing page counts)

• accuracy: as noted above, a page count gets incremented when the program is run, which is only when that URL is accessed by the browser. This is not the same as the number of times the user may think she has accessed the file. A browser may have a file in cache, and hence not reload it, or might conceivably fetch a file several times (e.g. each time you click on the "reload" button). IMG tag counters are of course not incremented when your page is visited using lynx. You might note that the various counters on this page do not show the same value!
• performance: server-side includes are rather expensive in terms of cpu time on the server. The Computing Center discourages the use of counter scripts that run on darkwing and gladstone.
• security: on darkwing/gladstone, the script that you write will run as you. If a malicious web surfer passes it data it doesn't expect, it could potentially do damage to your files.
• reliability: the system administrator may change the web server configuration in ways that break your counter. If you decide to make use of a counter service at another site, you'll probably find that the site eventually goes out of business and your counters break. Or performance accessing that remote site just to update your counter will be unacceptable.

Additional references

• Yahoo survey: Computers and Internet: Internet: World Wide Web: Programming: Access Counts