Laminates Under Pressure
By stern on Oct 10, 2005
Laminates and composite materials are the plastics of the current decade. Hockey sticks, auto body panels, golf clubs, and bullet proof vests benefit from a mix of kevlar, graphite, and other arrangements of long chain molecules. But if you want to mash-up an old school feel with a new school material, you end with what this picture used to be: a wood laminate blade on a rubberized grip graphite composite hockey stick. I captured the feel of a puck on a wooden stick through the high-pressure stacking of an exterior fiberglass grip, an outer shell sporting the perfect European curve to which Patrik Elias lent his name, and the inner core of something stiff (that's the technical term).
About an hour into last Sunday night's hockey practice, I was beat, and decided to hook my buddy Steve and have him tow me. I got my stick across his midsection, until someone else skated by using Steve as a pick and hit my laminated blade glove first. The ensuing "pop" had the sickening sound and fury of a carefully built balsa wood radio controlled airplane hitting a tree at a scaled Mach 2. What you see is what I got. It didn't just delaminate under pressure; the blade practically exploded.
Laminates exist in the software world, but we refer to the component pieces as having "sedimented" into another layer. That's a nice way of saying the pieces are glued together and we don't want to pay attention to the boundaries any longer. "But," you say, "aren't laminates strong?" Strong, yes. But they don't do well under sheering forces. As Jaron Lanier has pointed out for the past 8 years, software becomes brittle when too much of it sediments. That mash-up -- brittleness as a result of layering -- seems to contradict the common thinking around sedimentation. When the edges are glued up and neatly finished it becomes resistant to change. Apply significant pressure -- to upgrade, update, improve security, or integrate something else -- and the laminate tends to disintegrate in unpleasant ways.
Don't take this to mean that I'm opposed to integrated software stacks or assembled applications. Integration and assembly, particularly when done according to a software architecture with well-defined interfaces, are not lamination problems. Loud objections issue in response to "sedimentation", because I want to focus on the architectures for assembly, not the mechanisms by which products are compressed together. It's easier than cleaning up after a laminate is put under too much pressure.