tems or retail software applications--that have absolutely nothing to do with Web technology.
But here's the kicker: If you do these things, you will also be preparing your systems for
eventual porting to the Internet.
Make It Easier to Use
People don't compare your applications and products solely to your direct competitors' products. When you're designing an accounts-payable system, you can't judge its user interface against another accounting system. Instead, compare it to your remote control, your car, or your kid's Buzz Lightyear action figure.
People don't study products as they once did. Nobody reads manuals. Programs have to be as easy to use as
USA Today
is to read. You can complain about having to lower yourself to the standards that made
Short-Attention-Span Theater
a success--or you can quit whining and just build products that people will be able to use within 10 seconds.
Keep this in mind: If you make software, eventually it's going to get stuck on the Web. You know how distracting an environment that is, right? So make your program stand out by making it simple.
Make It Smaller
This is imperativ
e. Hard drive technology is only barely keeping pace with software bloat. My portable has a 687-MB hard drive--and only 20 MB free.
The discipline of smallness forces you to look carefully at all new features. Does your database system really need a singing and dancing company logo? Smallness also lowers distribution costs--especially on-line distribution. Finally, the smaller your program is, the more likely it will be installed on users' machines if they, like me, have a perennial space crunch on their systems.
My favorite example of this principle is Software Publishing's presentation graphics program, ASAP. It does about 75 percent of what Microsoft PowerPoint does. However, PowerPoint does about 300 percent more than the average presenter will ever need. And ASAP fits on two floppy disks. That's what I call progress.
Make It Faster
Yes, you can plan on Moore's law holding for the foreseeable future, but you can't plan on all your users or customers following the law
. Old PCs don't die, as you know. Instead, they get put on the desks of secretaries and assistants--people who spend countless hours every day running software. If you strive to make your programs faster, you're more likely to keep these heads-down workers more productive.
As corporations (and households) begin to adopt low-cost Web terminals, the speed of your code is going to become even more important. The $500 Web PC isn't going to have a 166-MHz Pentium in it.
Make It Cross-Platform
Windows is everywhere today, but it's not everything. Even Windows is changing--two years from now, Windows NT will be a much more common OS on business desktops (and portables). So don't limit yourself.
Fortunately, developing across platforms doesn't mean developing the same program a dozen different times. What it does mean is using the tools available to code to an intermediate standard, which can then be more quickly ported to other platforms. You must be ready to have your product
run not just on a high-powered Windows PC, but possibly also on a network computer (Web PC), or perhaps on Java running on a Unix workstation.
Am I asking for too much? Maybe. But if you make systems that are easier to use, smaller, faster, and multiplatform, you'll find yourself not just with better systems overall, but systems more likely to fit on the Web. That is, once you figure out your Internet strategy.
While You're Waiting for the Web
