If you missed the first post, start with 001. Becoming a Microsoftie (Chapter I). A Prologue has been added offering just a bit about my own history with computing. There is also a Roadmap/Table of Contents. If you need an overview and guide to subscriptions with discount codes, see Introducing "Hardcore Software".
Back to 001. Becoming a Microsoftie [Chapter I]
When I first arrived in Redmond, I lived about three blocks from campus, in company-provided temporary housing at an apartment complex called Bellevue Meadows, a block from the Residence Inn I had almost burned down during my interview. It was still light out at 9 p.m. on my first night (welcome to the Northwest), so I walked over to campus to check it out.
Microsoft’s three-year-old campus was made up of the original X-wing buildings, 1 through 6, and the recently completed double X-wing 8 and and 9. There was no building 7. Nobody knew exactly why, though there were a bunch of theories tossed around over the years. Sending a new person to meet up at building 7 was an ongoing prank. The real buildings surrounded a fountain and the small-but-infamous Lake Bill (which, looking back, is much smaller than I recalled from that first night). There was a basketball court near the lake, which always seemed to be in use. The buildings, connected by tree-lined sidewalks, were known for their design, which was meant to maximize the number of private offices with windows.
In a nod to Microsoft’s culture of self-reliance, the next day, my first day, after a two-hour orientation session (that felt like forever), I and about 20 other college hires were left to fend for ourselves. While I had been told how to set up direct deposit (paychecks were still hand delivered for many) and learned some details about my healthcare plan, I literally had no idea where to go. Fortunately, a more studious fellow new hire noticed, buried in the paperwork, that there was a map and an old-style printout with name (Steve Sinofsky), telephone extension (x67768), manager (Scott Randell, listed as SCOTTRA), and an email ID and password (yes, printed). The floor plan and numbering system made MIT’s infinite corridor look understandable.
As I flipped through the paperwork, I noticed my assigned email name was STEVESI on the printout, which immediately irked me as I was steven in all my previous systems (all lower case because of Unix). Obviously, I was not going to complain.
As I came to understand it, SteveSi was officially, and forever, my new name. Email names were how people wrote and spoke about others. Where a previous generation might have used only a last name out of casual respect or mister in person, Microsoft used email from BillG to SteveB on down. Aside from the free drinks, private offices, and khakis with button-downs, email names remained one of the iconic cultural identifiers of those days (and still used among alumni.) Given names no longer mattered at Microsoft. I was SteveSi. Cris Wittress was CrisWit—I finally got what she meant.
Before I started, Steven Schwartz had landed the name StevenS, a fact for which I was always jealous. After he left I even tried to secure the name, but there was a no-recycle policy. A coworker named Bill Gallagher was given BillGa, and for years he got crazy mail intended for (the real) BillG. As the company grew, it began to wrestle with the complexities of people getting married (or divorced) and how to deal with email name changes—much trickier than they’d imagine. Ultimately, in the late 1990s with the move to Microsoft’s email product, we finally moved to friendly names like email@example.com.
There was a list of email aliases (an early Unix-ism) to get help, like benefits, sickday, vacation, supply (office supplies), recept1 (recept2, recept3, etc. for the receptionists), stock (stock option sales), espp (employee stock purchase plan), payroll (for help with direct deposit), and, best of all, pcrepair, which could help with computer hardware. Perhaps that was second best, as I soon discovered library, which mailed the Microsoft librarians any topic to research or a request to send copies of articles or locate any book needed for work. There was an actual library filling most of one arm of an X in building 4, where I spent a lot of time as well. Everything was an email away.
Microsoft made about 35 different products back then, and I had personal experience with almost none of them. Importantly, by the mid-1980s, Microsoft moved beyond being a single-product company. It had substantial businesses in each of the major categories of the day: languages, operating systems, and applications. No single product represented more than half the company revenue. This early diversity was critical to Microsoft’s growth. In many ways, early software companies emulated record or book publishing by having many licensed titles for sale, and while early Microsoft followed this model it was now building most software in house.
The Systems group was the big group and was made up of the grown-ups. It felt to me the most like my summer aerospace job because there were people who were married (gasp!), and some even had children. This was the group that made MS-DOS, which was the single biggest moneymaker. They were also making OS/2, which was a massive joint project with IBM. There was a much smaller side project called Windows that was increasingly interesting. Unique to the Systems group was a much larger number of people who had joined Microsoft with years of prior work experience. There were people from IBM, DEC, ATT, HP, and a host of other computer companies from a previous era. Dave Cutler (DaveC), a legend with over 25 years of experience, had recently joined from DEC along with many of those colleagues. This made sense since building an operating system was something done at other big companies.
Languages was the history of the company and the oldest group. This was the group that made BASIC, as well as programming languages and tools from C to Pascal, Fortran, and, importantly, Assembler. The Languages products were for MS-DOS, Xenix (the commercial version of Unix, the ancestor of today’s Linux), and an expansion to OS/2 (an ill-fated joint development between Microsoft and IBM).
I thought many of the people I met in Languages seemed old. Some owned houses and had new cars. Some had been at Microsoft more than five years already.
Apps was the colloquial term for Applications, which is how the computer industry viewed programs used by end-users, versus the Operating System, which was required by the machine, or Languages used by developers. The Apps group was less tenured as it was both a newer business for Microsoft and seemed to have more college hires. Apps was almost a sleeper business even back then. Most of the products it made were for the Macintosh, like Word, Excel, and File, all of which were on the first or second version. Apps for MS-DOS were almost as numerous, but all were a distant number two in the market relative to software giants Lotus, WordPerfect, and Ashton-Tate, and Software Publishing that I had used in my summer job during college.
I walked over to building 5 to find the private, interior office in which I’d begin my career. It had no exterior window but had one to the hallway. As I searched for my office, I passed the kitchen and saw the giant glass-door refrigerators filled with cans of every variety of Coke and Pepsi products like a convenience store.
It would be decades before I paid for a beverage.
Just across from the kitchen was the mail and copy room. This room had everything one could imagine needing for work. It was like a CompUSA and Office Depot all in one. Along with a big laser printer (and a copy machine), there were 5.25-inch and 3.5-inch floppy disks by the case, notebooks of every size writing paper (not computers which hardly existed in portable form at this time), printer paper, pens, tape (transparent and masking), thumb tacks, and more. There were boxes of colored pencils (legend had it that BillG used those to annotate code with different colors, but I later learned that was a myth). There were rulers for scanning across lines of code in landscape. Best of all were the staplers with the Microsoft logo on them. This was like a gift shop, and anyone visiting left with a box of floppies and one of those staplers.
After a few wrong turns, I finally saw the engraved door placard (think Mad Men) that read STEVE SINOFSKY. Not Steven. I was peeved. While I did not meet him for a few years, Steve Ballmer (SteveB) had something to do with this, I’m certain.
Later that morning, I met a fellow college hire named Antoine Leblond, a French-Canadian who was in a far worse position than me, as the powers had reduced him to TonyL. That only lasted until his then-girlfriend visited and, as an even more ardent Québécois, Lucie Robitaille somehow managed to get it changed to a cool alias: Antoine.
Offices back then were furnished in what could be described as Native Northwest. Think a solid wood oak 60-by-30-inch deep desk with a 24-inch typing return and a swivel chair with matching oak arms. There was a matching 60-inch high solid oak bookcase. A whiteboard and cork board were attached to the white walls. A 12-button analog phone in corporate brown was on the return, featuring my personal phone number, 206-936-7768 or x67768. The furniture reminded me of the make it as indestructible as possible stuff that filled the freshman University Halls at Cornell. Even if I was motivated to rearrange the layout of my nine-by-twelve-foot space, I could not because everything was so heavy. The setup was also horribly non-ergonomic by today’s standards. Still, by any measure of an entry-level office, it was amazing.
My bookshelf was pre-populated with, I later learned, standard-issue books for every new software design engineer hire. There was an Intel 286 and 386 reference along with a Motorola 68000 reference—everyone in software engineering understood machine architecture and instruction sets. A phone-book-size MS-DOS encyclopedia weighed down the shelf. There was also a dictionary and thesaurus, and a copy of the same Microsoft Press desk calendar featuring important milestones in computing and an MS-DOS technical reference card in the back that CrisWit had sent as a recruiting gift.
Importantly, there were two seminal works on programming, Fred Brooks’s The Mythical Man-Month and Programming Pearls by Jon Bentley. The former, I learned, was the most epic of all Microsoft struggles, which was trying to release products on time—by the summer of 1989 Windows was on the second version, having shipped 1.0 almost two years after public announcement. The latter book represented the hardcore ethos of Microsoft software engineering, which was tight code—what code could be written to solve the problem with the most clarity and fewest lines, least amount of memory, and fewest CPU cycles.
There was also a copy of The Hacker’s Dictionary by Guy L. Steele, a famous computer scientist partly responsible for the programming language Scheme used and developed at MIT. The book was a 1980s version of what was often called computerese though Microsoft had its own unique language. One other book seemed rather strange to me, Stewart Brand’s The Whole Earth Catalog, which seemed useful if I was intent upon producing my own energy or building a yurt, but definitely represented the tail end of the hippie culture of computing from which we all originated.
There was a Compaq PC and a terminal in the office. The Compaq was an Intel 386 chip running at 33 MHz with an extended memory card and hard drive. The terminal was hardwired to Xenix servers via a different network and where the email system was hosted. It was Xenix email, which itself was just a port of Unix mail. I was right at home shelling out to “vi” to edit mail as I had been doing since college. (vi as in visual editor abbreviated.) There was also an HP-16C Computer Scientist calculator, for handling all the hexadecimal and binary conversions I would need to do, but I already owned one.
I signed on and changed my password to one I used for the next 10 years or so until password policies came into vogue. I fired off emails to some old lab mates who were the only other people I knew using email. I didn’t hear back right away, which was weird. That was when I learned outbound mail was batched and sent/received twice a day. I was told Gordon Letwin (GordonL), the legendary MS-DOS and OS/2 engineer, was not in favor of being connected to ARPANET or BITNET due to security concerns (he was certainly ahead of his time) so this was the compromise. Emphasizing this, our first business cards had only phone and fax numbers and TELEX numbers (!). By special request, a UUNET address could be added. UUNET was one of the first commercial internet provider of email addresses. That summer, mine was still uunet!uw-beaver!microsoft!stevesi. Don’t ask. Microsoft used email for everything inside the company, but externally email was not yet a thing.
Turning on the PC, I was immediately greeted with a hung machine (back then we called them machines, not devices) unable to make it through the boot sequence. I received my first lesson in corpnet, or the corporate network. The network was reliable, but the software on the PCs was not. Hangs were frequent and the only fix was a power cycle. I was only familiar with Novell Netware and had not yet experienced a product in the same space that Microsoft’s had just released, LanManager, a.k.a. LanMan. I wasn’t alone. Almost no one had bought the product because it mostly didn’t work.
This brought my first experience with emailing helpdesk. By email, they asked me if I was an SDE. I wasn’t sure, and then I realized my title was software design engineer. The next mail said they were on the way.
A nice man with a pushcart filled with tools and gear to keep PCs running and connected showed up. He pulled a 5.25-inch floppy out of a plastic disk holder and began the process of a network boot, which was a fancy way of using a floppy disk to boot a computer and connect to the network. After a minute or so of grinding floppy noise, I saw the magic “C:>” prompt.
The tech began some magic incantations that were new to me, like NET USE to connect to a shared network drive. Then he began to install OS/2 1.1 and then applications, but there weren’t many. I asked where the printer was and he laughed. I learned OS/2 didn’t really print yet, and to do so I best to use MS-DOS and those apps, which he then set up (also using some new magic like mapping LPT1 to the nearby printer in the copy room).
Once my computer was set up, I still wasn’t sure what to do with my day, but it was lunchtime. I was never really good at lunch or spontaneously meeting new people, so I began to get stressed. I finally resigned myself to passing on lunch and futzing in my office. Then I heard a knock on my door.
“Hi, my name is Andy Craze . . . AndrewCr.”
We were both new, though Andy had started the week before, and we were both joining Apps (the team would soon move to Development Tools in my first of many re-orgs). As recent grads do, we exchanged where we were from, college, and major information. Andy was from Cleveland. Went to Stanford. Studied computer science. He also informed me he was a huge Grateful Dead fan. He was outgoing and suggested we get lunch.
What I didn’t know, until Andy explained it, was that we weren’t technically working at our actual jobs yet, or even sitting with our teams. Instead, we were in Apps Developer College (ADC). ADC was where new Apps SDEs learned how to be Apps SDEs. We would be there for an indeterminate amount of time while we learned the ropes—meaning learned the tools and techniques of the Apps division. If those few sentences sounded like a bunch of jargon, that’s essentially what every conversation sounded like.
Unlike today’s start-ups in Silicon Valley, lunch was not free but marginally subsidized by Microsoft and operated by an institutional food company. We went to the pizza station and I ordered by the slice.
I sustained myself on pizza for a decade.
On to 003. Klunder College
Subscribers head over to substack.com by clicking on the title of this email and join in the comments and discussions. If you received this from a friend, please consider subscribing.