A few weeks ago, I tested an “off-line browsing” program. Such programs download pages from the Web unattended.

They provide a way to gather information from the Web without spending a lot of time; the program does the work for you. Then you can view the pages very quickly when it’s finished.

I’ve tested quite a few of these programs, and I’m not impressed with many. They seem to be overly complicated and don’t work too well in general. But this one was different; it worked quickly, did what it was supposed to do, and was easy to understand. There was just one problem, though. I couldn’t figure out how to get the program to grab Web pages from more than one Web domain at a time.

You see, these programs follow links from page to page, going “down” through the links as many levels as you specify. But this program did something a little odd. It wouldn’t follow links from one page to another if the link went to a different domain. For instance, if you tell the program to get Web pages from http://www.firstdomain.com/, and there are links from that domain to, say, http://www.seconddomain.com/, the program won’t grab the pages from that second domain.

Now, this seemed such an incredible weakness to me that I was sure there must be a setting somewhere to make the program cross domains, but I simply couldn’t find the setting anywhere. So I e-mailed the company’s technical support group. In return, I got a message saying that in fact there was no such setting; this product will only grab pages from one domain at a time.

That seemed quite ridiculous to me. How could the development team build in such an obvious flaw without realizing it. But more importantly, how could they build in such a flaw when their competition didn’t have this flaw! I sent e-mail back making this point, and received in response a message with several comments: a) they said that I wasn’t the first person to make this observation; b) they pointed out that if the product had this ability “you could be downloading much more than anticipated;” and c) they asked which competing products had this ability.

Now, I don’t want to go into the uses of off-line browsers, so just let me say that there are a lot of very good reasons why an off-line browser must have the ability to grab pages from more than one site at a time. More importantly, most off-line browser developer teams agree with me, as their products do let you do this. So how is it possible that this particular team missed such an essential feature? They clearly didn’t spend much time looking at their competitors’ products.

I’ve seen that this is the case with the development teams I’ve worked with, and I’ve also seen the products of many other teams that are released with glaring flaws that could exist only if the developers weren’t paying attention to their competitors’ products.

Here’s another example, the Netscape Navigator and Internet Explorer browsers.

In general, I prefer Internet Explorer, but that program has one weak feature in particular that really annoys me: the Find command.

In Explorer 3, the Find command doesn’t work very well for two reasons. It always searches from the top of the page, and you can’t scroll the page and keep the Find box open at the same time. So let’s say you search for a word and find it, but you’re not sure if you’ve found the right occurrence. You have to close the Find box, then scroll up and down so you can see the text around the word and determine if it’s the place in the document you’re looking for.

If it’s not the right one, you have to open the Find box and search again … but as it searches from the top (there’s a bug here, because it’s supposed to be able to search starting within the page), it will begin by finding the occurrence you’ve just looked at.

How about Navigator, though? That program’s Find feature always has worked well. It can search from any point within the Web page. You can leave the Find box open while you move the page a little to see if you’ve found the occurrence you want. And even if you close the Find box, you can press F3 to move to the next occurrence. (Explorer 4’s Find command is a little better than Explorer 3’s , but still not as efficient as Navigator’s.)

There are other weaknesses, too. For instance, in Navigator, you don’t have to type a full .com URL; instead of typing www.ncbr.com, all you need to type is ncbr. Navigator’s had that feature for a long time, so why hasn’t Explorer? It turned up in Explorer 4 recently, though it doesn’t work consistently. But why wasn’t it in Explorer 3? How does Microsoft miss such obvious weaknesses in their product? Because they’re not looking at their competitors’ products closely enough.

It always amazes me to see development teams who are not paying attention to what the competition is up to. Would a sports team turn off the TV during the season, and ignore what their competition is doing? Of course not, so why do software developers?

Here’s what I recommend, a fix for this problem.

Every development team needs to find someone who’s job is to keep an eye on a major competitor. That person may be in the team, or perhaps a technical support person or technical writer. Preferably one person per competing product. Now, I’m not suggesting that these people occasionally take a look at the competing product; they need to use it every day. They need to get comfortable with the product, to know it inside out, to know it so well that they understand its strengths and weaknesses.

There should be an important rule for these people; they’re not allowed to criticize the competitor’s product. It’s all too easy for these people to become infected by the development team’s cheerleading, a “we’re the best” attitude, and that’s not particularly helpful when you’re looking for ways in which your competition is doing a better job. Their job is to praise features, to point out where the competition is better. They should be encouraged to do this without fear of feeling disloyal; if everyone else understands that it’s their job to praise the competition, there should be no reason to regard these people as unfaithful.

I test hundreds of programs each year, and most have significant weaknesses. The sad thing is that most of these problems are easy to avoid, by learning the ways in which competing products do a better job. The information is there; it’s cheap, it’s easy to gather, and it can mean the difference between a mediocre product and a really great one.

Peter Kent is the author of Discover FrontPage 97 from IDG. He can be contacted at geek@arundel.com.