SCI Technology

May 1986 - July 1989

After graduating from UTC in May 1986, I started working at SCI Technology. SCI Technology was a division of the Fortune 500 company SCI Systems which has since merged with Sanmina to become Sanmina-SCI. SCI was a contract manufacturer who was known for manufacturing the original IBM PC and Apple Lisa. SCI Technology was a division dedicated to developing products which carried the company's own name. This division made ticket printers for United Airlines, SCADA, banking terminals, and parking lot systems. The group I worked for made Multibus I based mini-computers running AT&T System V or 386/ix.

While at SCI, I worked on a variety of projects but spent most of my time developing software for an intelligent serial communications board called the Data Communications Processor (DCP). The DCP was Multibus I with an 80186 processor, 4 Zilog Z8530 serial communication controllers and a parallel port. There was a similar board for the ISA bus. Up to eight of the Multibus I boards could be placed in a single system. I ported the System V TTY(e.g. termios), SXT shell layer, and LPT line printer drivers to this board. I did this by implementing a subset of the AT&T Device Driver Interface on top of the in-house RTOS IOPX (IO Processor eXecutive) which was supposed to be compatible with Intel's iRMX. I also spent time supporting ports of third party packages for SNA/SDLC and Bisync to support IBM 3270 terminals. The group I was in was quite small. I was also responsible for the diagnostics software for this board.

While at SCI, we did a project for Kodak called the Preview which could capture graphics from very expensive pre-press artwork systems and transmit it across leased lines to another Preview unit where it could be reviewed on highly accurate Barco monitor. The video card supported two frames of 512x512 24 bit color and I recall having fun developing the diagnostics for the video RAM boards. The hardware engineer helped me get the software to the point where it could tell him which pins where tied high, low, or not connected. Then he fixed them. The selling point of this device from Kodak's perspective was that it avoiding printing the image and shipping it to another site. SCI's point in doing this was to dump some 40MB (yes Megabyte) SCSI hard drives that were on the books at $1500 a piece. This device also used the same Multibus I infrastructure and helped our parts volume.

It was while working on the Kodak Preview that I got my first teaching opportunity. I taught a week long course on real-time programming and IOPX at the Kodak Hawk-Eye Factory in Rochester New York. Since the Preview board had LEDs, all examples we had the students develop used them in various ways. That concept is still found in the examples I use in the RTEMS Training Courses.

Ultimately, I left SCI because I was less than three years out of college and I was the senior software person. This left no one to mentor me and help me grow professionally. The company was geared to manufacturing and not flexible with its engineers. But the scary thing was that the group was clearly unprofitable and I worried it could be axed at any time. Strangely being at the tiny OAR was more comforting. At least I knew when we needed to find more work.