Communications of the ACM

• to count every visit;
• to count a visitor only once on each visited page (during a normal "stay" of 1 to 15 minutes), that is, avoid incrementing a counter a second and third time as one visitor, during "one visit," revisits a page; and
• to never count my visits.

I tried to contact someone about counter performance. The correspondents via email simply pointed to the online instructions. In July 2001 I wrote a letter to the chief executive officer of AT&T in the hope my letter would get to a customer relations office. It did. I received a phone call. In a few minutes the person had seen his own visits to my Web site using Explorer counted and had seen his colleague's visits to my Web site using Netscape not counted. A partial diagnosis was provided. The second of the two "cgi-bin" strings, the one produced by the Word 2000 HTML editor, recognizes and counts only visitors who are using Explorer. A partial diagnosis was provided. I was then referred to a chat room for further advice. I was told my "cgi-bin" string had to look exactly like the instructions prescribed and that then the Web site visit counter would count all visitors. Using Microsoft's Notepad, I opened the Word-created HTML file, revised the "cgi-bin" string to look like I was told it should look, and saved the file using Notepad. Having made this change, visits by users of both Internet Explorer and Netscape were being counted. However, visits by AOL users were not being counted. I later learned visits by visitors using current versions of AOL were being counted while visits by visitors using older versions were not.

I reported again the incomplete functionality of the counter, but got no response. The executive appeals office stopped listening to me. In October 2001 the counter management software in AT&T WorldNet's member services site was changed. Before the change, one could delete a counter no longer needed, in effect resetting it to zero. Under the new and current software, one can reset the counter to any value one wishes to have. Any Web site owner who wants to present visitors with big numbers can simply supply the big numbers. I have yet to receive an explanation about why the two "cgi-bin" strings work as they do, why the counter management software was changed, why customers are instructed to use one of the "cgi-bin" strings but not instructed to use the other, or why it is not possible to devise software meeting the specifications I have suggested—specifications likely to be useful to many customers.

In July 2001 I wrote the chief executive officer of Microsoft, explained the problem, and asked why the Word 2000 HTML editor does not permit a modification of its "cgi-bin" string. I have yet to receive an answer.

Illustrating another issue with respect to Internet use, the long technical article I published on my Web site has value to my fellow model airplane enthusiasts. Strangely, there is no direct equivalent for my posted article. My article can be read online. It shows my copyright. I ask that those who keep a copy of my article send $10 to the model flying club of which I am a member, a private nonprofit organization like many others, in exchange for taking a copy of my intellectual property as they might if it were printed on paper and available for purchase. As of this writing, the visit counter for my article shows 58 visitors, an undercount. Only one of those 58 visitors has sent $10 to our flying club. Over the same period, the visit counter on my home page shows 358 visitors, also an undercount. On the best evidence available, roughly one in six visitors in 11 months since starting my Web site viewed the article.

Many questions are prompted by this experience. An obvious action is to get the incidents and their underlying grounding published so they can be widely known and discussed. The following are a few of the many questions these incidents prompt, questions pointing to the need for further data-gathering as well as the exercise of considerable wisdom and the doing of much difficult work in changing what we in the computer science profession do with the computer technology at our fingertips.

Why does AT&T WorldNet place at least two Web site visit counters on its server but tell its retail customers about only one? Why does it put a Web site visit counter on its server that counts only visitors using Explorer? Who may see the counts accumulated by these counters—the retail customer only or other folks too? Why does Word 2000's HTML editor output only a "cgi-bin" string that points to software counting only visits by users of Explorer? Why is there no offer of explanation for the problem I experienced? Why was the software for Web site counter management revised in October 2001 so users can post any count they wish to show? Is allowing any count to be displayed consistent with social and business wisdom? Is it the customer's responsibility to test software? Is it the software provider's responsibility to repair a faulty program? Must small businesses, trying to start Web-based activities, hire computer experts to build customized visit-counting software before they can hope to get accurate counts of visitors to their Web sites? Have we generated a culture in which truth (in counting) and respect for intellectual property (as in downloading my article without having an obligation to pay) lost their meaning? Do we want an Internet world in which laws, regulations, and more investigators have to be created to catch and penalize those whose compliance with responsible practice and law is low? Are computer professionals willing to supervise their own on-the-job performance? What responsibilities do those in the computing professions have for the moral integrity of what they are doing?

Whatever became of integrity?