h a r r y @ h a r r y d o l e . c o m
History and adventures of Harry Dole follow herein.
Sr R&D Manager, Director, 2004 – current
Cadence Design Systems
Senior member of the consulting staff, 2001 – 2004
I did performance analysis on the Virtuoso product, including VXL. I managed a small farm of machines running 24×7 running performance test jobs. Managed in php, it has priority queues, jobs downloading from a mysql server, and errant jobs posting via email to folks that subscribing to various portions of the test hierarchy. I also worked on OpenAccess2.0 as it emerged, though had some difficulties with the extant architects in how multi-threading access should occur and how generic interfaces could be implemented with regards to OA2 and the current Virtuoso database. This harkens back to the days at Fujitsu where the plan was to implement tooling in terms of standard interfaces.
Fujitsu Microelectronics Inc
Chief Software Architect
Fujitsu Software Award 2001
Manager, Special Projects
An environment for designing integrated circuits. Computers include browsers for displaying pages of forms, with the computers in communication with a methodology server and a compute server. The methodology server contains design methodologies accessed by the computers, with the design methodologies defining steps of designing and testing of integrated circuits. The computers or methodology server are also in communication with a compute server. The compute server executes electronic design automation tools as requested.
Here’s a presentation I largely made that my boss David Dick gave regarding IPSymphony.
Fujitsu at the time made vast numbers of memory chips. A new format came along known as mpeg. I noted to the general manager one might use their memory chips instead of the emergent practice of writable CDs. After pausing he asked the very senior engineer next to him if that were real, and he paused.
Dazix, Veribest Electronics
This is in effect still Intergraph, or Intergraph Electronics. I transferred from the Clipper division to the electronics software division after they bought out Dazix. Eventually this became a vague spin-off called Veribest. My significant other at the time said it sounded like the name of a dairy.
Sr R&D Manager
My time spent teaching calculus came off well here as a lot of time is spent talking to clients and giving presentations.
But the more economically significant part is assisting in winning the NAVAIR/SPAWAR CAD contract in the early 90’s.
CAD manager, 1987 – 1991
Intergraph bought the Clipper microprocessor team when Fairchild went under, so it was a smooth transition to Intergraph.
Tangent Data Systems: Adapted and worked with Tangate and the early horrors of LEF/DEF. Intergraph sold Tangent to Cadence Design Systems, so somehow I got the job to go into Cadence’s offices to port Virtuoso to the Clipper platform to promote more folks to buy Intergraph machines.
I managed the CAD production of the Clipper C3 and C4 microprocessors. The C4 had a lot of floorplanning techniques incorporated into the the Interconnect Description Language (IDL) that the gate-slingers use to hook up the standard cells. Essentially this was an easily described clustering method in the high-level language for the interconnects.
VAX VMS-780 and Ultrix until the Clipper C100’s arrived in cards for slipping into desktop IBM PC clones.
Hung a lot of Clipper Versatec plots on the walls. One compiler worker was having some difficulty with the barrel shifter, and I pointed out to him it was right outside his door. He looked at me oddly as I showed him the reasonably simple data path on the wall.
CAD engineer, manager, 1984 – 1987
A friend of mine invited me to work in Silicon Valley as a contractor to work on the Clipper microprocessor. This seemed like a golden gate opportunity, and it stuck as it caused considerable learning. My actual first machine was one that I tried to design in the 4th grade using the only vague thing I could think of: delay lines. Could not figure out what a transistor was though, so it was hopeless.
The article is about how to use simple pieces to build complex things. Main ingredient is category theory.
Incremental Static Timing Analyzer
Fun with Fortran this was. In those days we had gate-slingers. These days we have engineers write in a so-called behavioral language and have it compiled, or synthesized, down to gates. The gate-slingers instead do this directly. This static timing analyzer allows the gate-slinger to rearrange the netlist, or the connections between the gates, and then incrementally determine the new timing for the propagated circuit.
Contributed to C Behavioral Programming Language
Versatec plotter drivers
So I plotted the chip at billboard sizes and hung the strips all around the hallways. As EDA software those days was somewhat wonky as few were attempted what we did, and vt100 terminals provide not the best imagery, I’d stare at the plots to look for LVS (layout vs. schematic) and ERC (electrical rule checks) errors. About that time a frustrated compiler guru walks out of his office and into me staring at the plot. He whines about some lack of feature in the barrel shifter, and I say, “there it is,” lurking there just outside his office. Epiphany it was for him to connect the abstract object with the physical manifestation.
University of Texas San Antonio
Assistant professor of mathematics, 1981 – 1983
Taught calculus, linear algebra, typical undergraduate math stuff. I recall my business math 101 class having 99 students where the threshold for a teacher’s assistant is 100. When I tried to get my prof’s parking permit I had trouble at the desk as the attending woman did not seem to believe I have the appearance of a prof. Need haircut.
University of Chicago, Mathematics, PhD., 1981
Being the only place that accepting me, out of Harvard and Princeton and Chicago, I was off to study under Saunders Mac Lane as one of the two co-founders of Category Theory. Luckily he took me in as a student. I thank Ed Oxford at USM for that insight to go to Chicago. Learning category theory changes the way you think. Here’s my reference in the Math Genealogy Project.
University of Southern Mississippi, BA, 1975
Getting a BA in mathematics is a bit unusual. I discover about 2.5 years into it that I have enough hours to graduate, but not enough to cover a science minor. I have plenty of hours in languages (French and German) as I need them for the PhD language exams. So the USM Dean grants me a BA instead of a BS in mathematics.
I took one Fortran class and only showed up for tests. B. I pretty much avoided machines in this era.
I was lucky enough to see and feel Apollo XVII do its night launch for the last hominid moon trip. The 2-3Hz waves invoke considerable frisson.
Harry Dole Dodge
I started pumping gas the same summer as hurricane Camille. With the power out, Dad had me out in back of the gas station with a hand pump lowered into the tank, with me turning the metal crank to pump a cup per turn wondering if it could spark. Not the last bit of fun with Dad and pumps. He put a bilge pump in the pool to drain it, but used a worn extension cord to power it. When I went barefooted to look what was going on, I was wondering why my leg was tingling till I noticed I was standing in a puddle with an exposed wire. Luckily I also survived the Vienna sausages from the gas station.
This portion of my work life shows up on my social security statement. Dad was assiduous in doing social security forms, particularly those he picked up in Palmer’s Crossing to go out into the woods and poison shrubbery under the pines he was farming.
“And about that time a bevy of bird dog done flew overhead.”
I memorized a large number of seven digit Chrysler part numbers. Toyota had two hyphenated groups of five digits. I prefer seven as that seems more than enough though others prefer ten as a 5×5 group/sub-group.
Wagon Wheel Motel
I delivered ice for tips. And molded dams in the gravel of the driveway after thunderstorms.
At about nine I wandered off alone into the Tatum Woods south of the Wheel. No video games so why not woods? I had found a still there before so onward.
There were peculiar mounds of earth in the woods. And crescent shaped buckets attached to pine trees just under a bark-hacked chevron just splendid for dripping concentrated pine fuel for ovens dug into the embankments.
But I digress.
Following the mounds into the woods, they were the remnants of a wood-railed steam train. Probably not Choctaw. And they led somewhere. Walking along the line I could see a lake emerging from the terminus. The logs stopped here and floated off to the right into the giant shadow of the looming great Tatum Saw Mill.
Eyes widened as I entered the vast orgy of rusted mechanics spread over two floors of the grand hall’s conveyors and saws capable of reducing original growth to sawdust.
A few years later I tried taking Roger out there in my Dodged Colt one evening and someone let a round loose as we entered the grounds. We skedaddled.
There is an article for that.
1970: Olivetti Programma 101
1971: IBM 360/40
The blinking lights! The big emergency pull button! The chain printers!
I still have a crate full of 360 manuals that I asked dear parents to buy for me when I was in high school. These are the sort that one would see mounted on a metal frame next the Selectric that served as a real console in the day.
1982: IBM 370
1983: Zenith Z100
1984: VAX/VMS 780
1985: VAX/Ultrix 780
1986: Clipper microprocessor
and so on