For example, I needed something that would let them decide when to break a page and when to not break a page (in other words, split an article into two or more parts, and let the editor decide where that split was to occur). I needed a tool that would help them decide what would be a feature article and what wouldn't. These are all editorial-level decisions and shouldn't be hampered by the need to do programming to produce an issue of the publication.

What we needed was a tool that editors could communicate with on an editorial decision-making level, as opposed to a tool that would communicate with editors as though we were (or needed to be) programmers. We found nothing that would do that. In the spring of 1997, it became apparent that we were going to need to build something. This "thing" became ZENPRESS, the tool we use to produce our publications.

We'd been publishing articles about AvantGo in PalmPower and Windows CE Power on and off for a while. But I really didn't think all that many people would want to read a magazine on a little tiny screen. But one day, we were going over our Web server logs and we decided to check out which browsers were used most. It turned out that AvantGo was used 23% of the time to read PalmPower pages (it was much less for the other two publications). But since we didn't have an optimized version of PalmPower, 23% of our readers were reading a poorly formatted, hacked together version of the magazine.

We decided to extend ZENPRESS (which, effective in early August, achieved "patent-pending" status) to generate handheld content. We figured we'd build something that the "die hards" could use to read on their handhelds and we'd get some experience and probably an article (this one, in fact!) out of it.

Here's where it gets interesting. A Web page built for a desktop browser that might support 800 dots by 600 dots that can each be one of 16.7 million colors is obviously going to need to look substantially different on a handheld device that's 160 dots by 160 dots and black and white. Yes, I know that many Windows CE-based devices support color, but we had to design for least common denominator if we wanted everyone to be able to read our publications.

Also, while you might have a 20 gigabyte hard drive on your desktop, you might only have a total of one megabyte of memory on your handheld. As a result, we needed to create Web pages for handheld systems that look and feel substantially different than those that would be viewed on a desktop computer.

In particular, we had to reduce images to black and white or eliminate them entirely. Eliminate the use of color. Perhaps use shorter text. And, if readers were going to be using something like AvantGo where they'd be mobile and not connected to a network, we'd need to make hypertext links respond in a different way.

Since readers wouldn't be wired to the Web anymore, we would need to restructure our links so that when a reader clicks on something, it takes him or her to an internal piece of information instead of trying to look something up on the Web.