As you may have noticed from the home page of Lionel Deimel’s Farrago, I maintain the site with Microsoft’s FrontPage 2000. FrontPage is a Web design and authoring tool with many virtues, though it seems to have a mind of its own that makes certain sophisticated tasks harder than you think they should be, rather than easier. Nonetheless, FrontPage is a great tool for developing sites to be maintained by others who have limited Web-development skills or for developing sites that don’t have to live on the bleeding edge of Web technology.
FrontPage accomplishes some of its fancier tricks—generating tables of contents automatically, for example—by means of special software installed on the Web server, known as FrontPage server extensions. FrontPage server extensions are widely, though not universally, available on Web servers on a variety of platforms. The latest version of FrontPage is FrontPage 2002, and there are server extensions that correspond to this product, just as there were server extensions for previous one, FrontPage 2000. The 2002 server extensions also support FrontPage 2000.
I especially appreciate a program like FrontPage for the automatic changes it makes for the author, such as adding links to new pages on the site map page. Another nifty feature I have made use of on many sites is the ability to display the date a page was last edited without having to change the date manually. This feature was very nicely implemented. Frequently, one wants to put a standard footer on some or all of a site’s pages, and this is an attractive place also to display the last-edited date. In FrontPage, one can place independently defined borders on any edge of the page and have it appear, say, on all pages. If I want to change this border, I do so in one place, and that edit will be reflected on all the pages seen by the Web visitor. Moreover, if I put a date in a border—the bottom border, say—even though the “same” border is used on every page, the date shown can be made to display the date the page itself was last edited. I used this feature to show the edit date of most of the pages on my Web site.
Unfortunately, a few months ago, the feature stopped working properly. When I updated a page, the date in the bottom border did not change. It only changed if I changed the border itself, in which case, all the pages with the border would show the same updated timestamp. Actually, it took some time to realize what was happening, or, more properly, not happening. What seemed especially curious was the fact that the edit dates of my pages were being displayed properly on my site map page—FrontPage indeed knew when the pages had been changed.
I first noticed the problem in late September or early October. I was busy with a number of projects, however, and did not get around to serious troubleshooting for a while. By early December, however, I had tried everything I could think of to diagnose and fix the problem, so I called technical support at SBC WebHosting.com, which hosts my site. The ever-helpful SBC people had a number of suggestions, but no one seemed to have seen this problem before and, after a number of conversations, it was not clear whether we were closing in on a solution, or even a diagnosis, for that matter. In parallel, I had opened up a dialog with Microsoft support on the Web. Much of this dialog was carried on in written form, but, over a period of about a month—I took some troubleshooting time off for Christmas—I also had several telephone conversations with the Microsoft person working on my case. I also searched elsewhere—though not everywhere—for insight into the problem. I looked at on-line help, books on FrontPage, and the Microsoft Knowledge Base. What I discovered was that the feature I was using was not well documented, and failures of it seemed not to be documented at all.
By mid-December, all the technical support people were telling me that there seemed to be nothing wrong with my Web site or the FrontPage extensions on its Web server, and everything seemed to work fine for them. (I have reason to doubt this last assertion.) In fact, when I loaded my site onto the Web server on my computer, it worked fine. Evidence that perhaps I wasn’t crazy then came in the form of an e-mail message from a client experiencing the same problem with another site hosted by SBC. I called him immediately to tell him that he wasn’t crazy either and that I was working on the problem. Interestingly, the client had noticed the problem about the same time I did.
When I got back to work after Christmas, I started asking the SBC people what had changed on my server. I eventually learned that the FrontPage 2002 server extensions had replaced the 2000 server extensions sometime in August. This was suspicious. Microsoft suggested I sign up for a free Web site that supported FrontPage and put up a few pages to demonstrate that the date functionality was or was not correct. I did so, and I encountered the same problem I was having on the SBC server. At this point, SBC was suggesting that I should perhaps be on a Windows 2000, rather than a Unix server, and that perhaps FrontPage 2002 would fix the problem. My Microsoft contact, meanwhile, gave up and handed off the problem to the next higher level of technical support.
I ordered a trial copy of FrontPage 2002 and began looking into how much more expensive it would be to host my site on a Windows 2000 server. A couple of days later, I got a call from another Microsoft technician. We had a long and interesting conversation, but what he called to tell me was short and sweet—that the feature I was trying to make work had been quietly eliminated in the FrontPage 2002 extensions. The “date this page was last edited” option of the time and date component apparently used a lot of CPU time on Web servers when pages were uploaded, and Microsoft was urged by Web hosters to eliminate it. Microsoft complied and apparently hasn’t heard much about it from FrontPage users. I wasn’t happy to hear this, but I was happy to know that I had no more need to beat my head against the wall. I know when I’m beat. I have modified my site and will send a request to Microsoft to restore the deleted feature. I will not hold my breath.
Although I thought the feature I had lost was quite useful, it was obvious that it was poorly understood and little-used; Microsoft could certainly justify their eliminating it. On the other hand, the feature might have been more popular had it been better documented. I do miss the days when software was documented in reference manuals that told everything anyone could possibly want to know about a program, including its behavior in every conceivable circumstance. Alas, software now seems to change faster than it can be tested or documented. Instead of reading reference manuals, we read about bug fixes and workarounds.