John Chamberlain
Developer Diary
 Developer Diary · You Heard It Here First · 23 November 2003
Palm versus PocketPC - the Final Analysis
Thank god the PocketPC is finally dying. This ill-advised exercise in hubris has been losing money for years. Any other company without Microsoft's vast pool of cash would have thrown in the towell or been bought out long ago. Instead the reputation and resources of the world's largest software company have saddled untold numbers of users with an inferior product and distracted the efforts of thousands of smaller developers on a doomed platform. What can we learn from this? I think there are two lessons here: good design is more important than business as usual, and what constitutes a good OS design.

Good Design is More Important Than Business as Usual

Many of Microsoft's initiatives have had the air of pride. It did not matter if it was good or bad, it was Microsoft so it would succeed. This philosophy has repeatedly been shown false. For example, with failed products like Tax Saver and perennial second-rate products like Microsoft Money. Nevertheless the business-oriented sales people who tend to take over any large corporation have the idea that better business is the solution to any problem. The ultimate proponent of this idea in the computer industry was Compaq Computer (now non-existent). It is no surprise that when Microsoft first developed PocketPC their first "partner" was Compaq. Compaq then used every business trick in the book to push their PocketPC-based devices: rebates, volume deals, vertical deals, you name it. If it would inflate sales, real or imagined, Compaq would do it. The consequence of this was that tens of thousands of middle management corporate types all over the country ended up being given Compaq Pocket PCs which they promptly stashed in their desk drawers. Even today there are hundreds of thousands of old Compaq Pocket PCs sitting unused in desk drawers. But hey, it was still a "sale", right? Meanwhile people who bought with Palms with their own (not their company's) money actually used them.

Now Compaq is a goner, but its iPaq straggles on under HP's checkbook. HP's personal division lost almost $400 million last year. Not all OEMs have the same patience and deep pockets as HP. One vendor after another has lost money on the PocketPC and dropped licensing. As things got worse Microsoft went grovelling with wallet in hand before obscure manufacturers like Acer, a Taiwanese knockoff artist and PC clone has-been. Ultimately no matter how much money and business power you have you cannot make an inferior product into a success (unless its Windows I guess).

What Makes PalmOS Such a Superior Design: No Saving

The key to the Palm OS is that there is no "saving". In typical software, all windows-based software for example, the user must explicitly give a "save" command to commit their work to disk. This allows the program to only occasionally write files instead of re-writing them constantly after every small change. Jeff Hawkins, the inventor of Palm, realized it would be much better to force software to save automatically and constantly rather than put the saving burden on the user. Consequently he designed the whole OS around this concept of constant and invisible-to-the-user saving. This is what makes the Palm so eminently usable.

Of course there are other factors, but the no-save factor is the most important. Windows and most of the software written for Windows is based on explicit saves so they could never emulate this behaviour of the Palm.

The no-save design has many effects. For example, more screen real estate is available because there are no save buttons. Another factor is that users can work faster because they do not need to take time out to do a save action. Another factor is that work is never lost due to failure to save or crashes--a common problem with Windows and other operating systems.

This brilliant innovation by Jeff Hawkins has had a huge impact on human productivity and will likely influence the development of future operating systems.

Developer Diary · · bio · Revised 23 November 2003 · Pure Content