For those that have an interest in knowing the gory details of how this works, I wrote a shell script called wlog, which I call, passing the URL that I am interested in logging. The script, in turn, calls a C program that I wrote called webtitle. This program visits the web page and extracts the <title> information, which is passed back to the shell script.
The shell script then creates puts two entries at the top of a log file, wlog.txt: the time and date of the entry, and a html formated line containing the entry itself.
When this web page (wlog.shtml) is called, a cgi script, wlog.cgi is executed. It iterates through the wlog.txt file and formats the entries according to date.
The only thing that doesn't work right now is that pages that respond only to http/1.0 queries return a (null) for the page title. This is because my C code isn't yet smart enough to satisfy the requirements of http/1.0. Right now it isn't that big of a problem because there aren't many pages that enforce it: CNN and Amazon are the only two that I've encountered, though I've not gone fishing for others.
