“Frankly Lou, I have had summer jobs that have been much more interesting,” I told my boss.
Running enthusiastically into my boss’s boss’s office I blurted out “Bob, I heard this engineer Franco describe a test method called Weighted Random Patterns. I found it so fascinating.”
As a young engineer at IBM in the mid-1980’s I made the above two statements. It could have led to nothing, instead it led to me working on a Sexy Hard Problem.
I had been at my full time job at IBM for less than a year. However, work bored me; I felt unchallenged. While studying for a Masters alleviated some of the boredom it didn’t address my motivation to show up at the office. I required more interesting work. In a discussion with my manager, Lou Caprio, I didn’t hesitate telling him so. During my summer engineering jobs, I had worked on more interesting engineering problems which provided a challenge. We discussed who I could work with next and on what—probably another gate-array circuit building block evaluation (reference previous blog post).
IBM had an extensive training program for their staff; and I recall being encouraged to take internal courses that included statistics and bi-polar circuit design. The training department offered a 10-12 week course on integrated circuit testing. Each week a different speaker presented; I found the subject interesting— defects occurred and testing detected them. One week Franco Motika explained Weighted Random Pattern testing; I left the classroom excited. I just had to share my excitement with someone; when I returned to the office I could not find Lou. Bob though, Lou’s manager, was in his office. Hence, he became the receiver of my enthusiastic acclamation; he smiled.
Those two statements I probably made within a few weeks of each other. Lou mentioned to Bob that I needed something more challenging. Sometime later, a sister organization came to Bob and asked for help with evaluating Weighted Random Pattern Testing (WRPT). Bob remembered me- bored with work and excited by WRP. Lou offered me the opportunity, I took it as I knew I would be working on A Sexy Hard Problem. Engineers thrive on such projects.
Semiconductor manufacturers use Automatic Test Equipment (ATE) to screen Integrated Circuits (ICs) for defects. ATEs apply patterns of 1s and 0s to an ICs’ pins; these patterns had to be loaded in the ATE’s memory. Complexity of the IC drives the number of patterns which in turn increased the time to load the patterns. In the semiconductor testing world, time impacts capacity of a test factory floor. What if patterns could be generated at the point of contact between the ATE and the IC? The engineers who conceived of WRPT were motivated by solving this very hard problem; their solution a creative approach which built upon ideas discussed by test professionals.
Evaluating the first WRPT application challenged me and exposed me to creative engineers who in turn inspired me. By expressing enthusiasm for the topic and boredom for the initial assignments my management team offered me such an opportunity.
Have a productive day,
Anne Meixner
Dear Reader, What memory or question does this piece spark in you? Have you been bored at work and raised it as an issue to your boss? Has a talk inspired you to want to work in that area? Please share your comments or stories below. You too can write for the Engineers’ Daughter- See Contribute for more Information.
Additional Reading
Information on Automatic Test Equipment.
Weighted Random Pattern Testing is based upon self generation of patterns. Linear Feedback Shift Register (LFSR) represents a common means of generating patterns.
Numerous patents exist for this technology, here’s one of many.
Publications on WRPT typically require fee-based access. If you have access to a college-level library free access is likely. I recommend starting with the IBM Journal of Research and Development paper; “Method for Generating Weighted Random Test Patterns.” A direct link is not possible, follow these steps:
- go to IEEE Explorer
- type “Weighted Random Test Patterns” in the search field
- select Single Year in the Refine Results widget
- type 1989
Top paper in the results is the one I’m referring to.