I’ve tried to keep the number of obsolete reference manuals and technical books I have to a minimum over the years. That stuff has been getting outdated at the same rapid rate as the evolution of the technology industry. And with on-line references available for all things technology-related, there is almost no need to keep paper copies of anything.
Despite best intentions, however, possessions tend to accumulate, and when we moved from Seattle to New York a few years ago after being in the same house for close to two decades, it was necessary to do some significant culling. If I had a book or manual that didn’t pass the “will you ever use this again” question, it went into the donation pile. The Friends of the Seattle Public Library organizes book sales every year to support the library, and this made saying goodbye to about thirty boxes of books our family assembled much easier. In this process, I did make allowances for sentimental reasons.
One of the exceptions I made was to hang on to my original copies of Borland Turbo Pascal. It came on a single 5.25” floppy disc along with a paperback reference manual. This is a picture of the original 1.0 and 2.0 versions that I’ve kept:
I credit this product as much as any other for taking me down a path that would lead me to become a professional software developer.
I was an undergraduate at Middlebury College when I bought it. Much of the software development I was doing was self-taught using one of the earliest IMB PC clones available – a Sanyo MBC-555. The Sanyo was not a very good machine and had lots of problems with compatibility, but it was the cheapest PC I could convince my parents to buy.
I had reached the limits of what I could do with Basic, and let’s face it – a real program was a compiled, self-contained executable package (a proper “app” for all the young readers out there), not some Basic file that you had to run through a slow interpreter. Also, I had been involved with assembly-level programming since the beginning of my interest in computers, and wanted a tool that allowed access to BIOS- and hardware-level functionality, even if it meant hand-compiling the opcodes using the 8086 CPU reference manual.
Turbo Pascal would let me do all of this, and at a price that a college student could justify to his parents: $49.95. This was a bargain compared to the high cost of any of the Microsoft tools available then; Microsoft’s Pascal compiler was $400. That was a lot of money back in the early 80’s, and a $400 compiler for a student was out of the question. At the time, I couldn’t have imagined that I would eventually go work for the Microsoft that wanted so much money for a software development tool.
I bought Turbo Pascal mail order, sight unseen. There was no Internet as we know it today, no Amazon, no on-line reviews, and my connectivity consisted of a 300 baud modem (that translates to 0.00029 megabits). Everything I knew about the product was contained in a glossy advertisement in Byte Magazine. I realize how quaint that all sounds, but when I got the package with the small paperback reference manual and the floppy, I was in programming heaven. The compiler was incredibly fast even by today’s standards, and produced real executable programs even if they were limited to the smaller .com variant rather then .exe files. And the fact that Middlebury’s math department taught a few Pascal classes (the college did not have a computer science department back then) was a big plus.
I would remain a big Turbo Pascal fan for a number of years until I fell in love with the C programming language, but that’s another story that also involves a thin paperback that I have also kept to this day.