Big Defense Firm: Weak Computers
More absurdities from my job at Big Defense Firm.
When I started I didn’t get a computer right away. It took almost a week to finally get a machine. While waiting I read documentation and shared computers with my teammates; we were all contractors; only our manager was a full-time employee.
When I finally got my computer, it turned out to be an outdated and underpowered machine. It was a Pentium 3 with 256MB of memory, which was quite bad for 2008. The processor had been discontinued four years earlier, an eternity in those days: Moore’s law was still going strong and every couple of years processor speeds doubled. My Pentium 3 was about four to six times slower than a new mid-range processor from the time. Memory-wise, 256MB wasn’t much; it was enough to run Windows NT 4 with Visual Studio 6, but anything else would slow the computer to a crawl.
It wasn’t just me who had an underpowered workstation; this was the standard configuration for most developers in the department.
There was so little memory on our workstations that we couldn’t run everything we needed at the same time. The application we developed depended on a database; the problem was that we couldn’t run the tool and the database alongside Visual Studio 6. There wasn’t enough memory to run everything. If we tried, the system swapped like crazy and was almost unusable.
When we wanted to test our code we had to:
- Shut down the database and close the application.
- Open Visual Studio 6 and launch a build.
- Wait at least 15 minutes for the build, with the linking alone taking over 10 minutes.
- Close Visual Studio 6.
- Start the database and wait one minute for it to be operational.
- Start the application.
Testing a simple change took around 30 minutes. Rebuilding the entire codebase took almost half a day on our antiquated computers.
Later I befriended folks from another department. They told me that their own equipment was more modern, and that their IT teams kept old RAM sticks and that I could get some if I wanted. We went there together and the cheery IT guys were glad to give me two 64MB sticks of RAM. I promptly installed the sticks into my station; I now had a whopping 384MB of memory, more than most of my teammates. I could finally build the software without shutting down the database.
It seemed like complete madness not to upgrade the hardware to something more modern. A new computer would have cost about two days of what the firm paid to employ me. Even if the new hardware made us only 10% more productive, the investment would have paid for itself in about a month.
I brought this up to my manager, but there was nothing he could do; he used the same clunker as everybody else and wished he could get something better. I brought that up with the head of the department during my 6-month review. He acknowledged that these machines weren’t great, but he couldn’t justify the investment with the bean counters who apparently were opposed to any new hardware purchase. He argued that these underpowered machines were actually a good thing: they made developers more thoughtful; the friction from the slow environment forced them to think harder about what they were doing. This sounded like lunacy to me.
Since this was isolated to our department, I assume this was some kind of scheme to keep the costs down. I imagine some execs and accountants tried to look good by showing their bosses low hardware costs for the past three years.
Stranger still: every month the department added more contractors to our project to compensate for the additional delays. After a while the Big Defense Firm ran out of obsolete workstations, but they were still adding more people to the project. That’s when they decided that new contractors had to supply their own computers when starting.
For three months I watched with envy these new hires roll in with modern machines that could build the software four to six times faster than my sputtering Pentium 3. I left the Big Defense Firm as soon as I could and never looked back.