Monday, January 25, 2010

Technology Fads Come and Go, Architecture Endures

Let me state for the record, and I hope my kids read this, I have never ever worn a pocket protector. I won't even wear one on Halloween. I discovered once that the growing cloudy smudge on my computer screen was actually the reflection of an expanding ink spot emanating from the uncapped felt tip marker in my shirt pocket. Still, I refuse to use pocket protectors.

And yet, with the pride of accomplishment one generally reserves for marathons, holes-in-one, or setting the VCR clock, I proclaim that I am a geek. Yep, an unabashed, unapologetic, unprotected pocket and all, card carrying geek. I have not one, but two fully functional Apple Newtons to prove it.

There is a phrase clothes designers use; Never confuse Fad with Fashion. For five years I managed my company's Advanced Technology Group, and one thing I learned is that even though I love, breathe, and live for new technology, it is important to remember that technology fads come and go, but architecture endures.

In 1992 a company called PointCast emerged out of California with a technology and business model that looked terrific. Newsweek proclaimed "PointCast would have to be aggressively stupid to fail." It combined push technology, distributed Internet servers, advertising revenue, and rich content to provide valuable information on user's desktops at no cost. Their software even accommodated an internal / intranet 'channel.' Success was guaranteed.

My team looked closely at PointCast and even established a small internal pilot to study it closer. One early decision we made was to store all of our internal content in Lotus Notes and build an XML-based bridge to PointCast rather than using the PointCast repository to hold our news and information.

We loved PointCast and for a while was on their council of advisers. We thought for sure that advertising would be a viable vehicle for providing free Internet services (long before Google). But we also knew that if we ignored sound architecture concepts (loose coupling, open API's high cohesion, etc) we could find ourselves locked into a product rather than a solution.

Needless to say, our dedication to sound architecture paid off, and while PointCast died as rapidly as it ascended before the end of the 1990's, our Notes-based repository including its XML-based interface is still in use.

The point is that you just never know what will and will not survive. Ethernet and the Internet were both doomed to fail as networks grew; and yet 20+ years after the experts predicted they're both thriving. If Apple's Newton ever makes a comeback, I'm ready.

No comments:

Post a Comment

Follow by Email