You may want to reconsider before you step into the time machine. First of all, 1991 in particular is a tough year -- it was a crushing recession in the US, and young people were having a really hard time finding jobs. So you wouldn't be "heading to the recruitment stalls" for any of those companies except potentially Microsoft. And this is absolutely the "why-are-manhole-covers-round" era of Microsoft, and it's the DOS era as well -- so not only is the company smarmy, its products are buggy, demoralizing piles of death-marched junk. Read the (excellent) Showstopper! for a hint of what awaits you at Microsoft.

There's a big difference between 1991 and just a few years later of course, but even when I graduated from college (1996), Microsoft was absolutely suffocating. I had decided that I wanted to work for a computer company and that I had zero interest in working on Windows NT(or Copland). This left one company, Sun Microsystems, which even in 1996 was not really recruiting at universities. I got a job there by cold e-mailing a Sun engineer (Jeff Bonwick) based on a Usenet post in comp.unix.solaris. (Cold e-mailing to get a job was so unusual that a friend of mine who was a reporter for the AP wrote a story about my job search -- and it was broadly picked up nationally![0]) At Sun, I was the youngest person in OS development by a decade, and the industry broadly thought Sun to be foolish for insisting on innovating in the operating system. Conventional wisdom was wrong, of course, and I had a great 14-year run at Sun that I wouldn't trade for anything -- but it would be a mistake to overly romanticize what was honestly a pretty crappy era.

[0] I talked about this briefly in https://www.youtube.com/watch?v=IznEq2Uf2xk, including an (embarrassing) photo of me ca. 1996 that ran as the front page of many business sections around the US

It's great to get this extra perspective from you Bryan, because you're probably the person whose talks have most made me interested in seeing 1990s Sun Microsystems et al. Point taken about 1991's struggles. I was vaguely aware that it was a troubled time, but took the premise as is. Which particular part of the 90s was best seems an awkward question. Were things generally getting better through the 90s until dot-com bust turned everything to shit?

Thanks for the book recommendation. I've added it to my Goodreads list. I'll checkout the talk too. I haven't seen that one of yours, because I saw the title and figured it was heavy on memory-management and pretty out of my wheelhouse.

It's hard to know -- but something that definitely struck me was that we at Sun did much better technical work in the bust (2001-2005) than the boom (1997-2001). (It's hard to say anything definitive about this as there is so much there that's specific to Sun and the group of technologists I was working with, but it certainly didn't make me long for the boom!)

As for the progression of software development, this may be trite to say, but in my opinion the single most important revolution in software is the rise of open source. I don't think that there was a single event here per se but rather a multi-decade long progression across many different domains that has changed just about every single aspect of software development. (Cloud computing too, certainly -- but I don't think that cloud computing is economically viable without open source.) So software development from the 1990s is unrecognizable in large part because it is just so damned proprietary -- which, as it turns out, was in fact a relatively short blip in the fullness of history.[0]

[0] https://www.youtube.com/watch?v=Zpnncakrelk, in which I expand a bit on the history of open source, among other things

I'm early enough my career to not at all find it trite. Open-source has certainly been the most exciting and fulfilling aspect of my software work so far. If I was thrown back to the 90s with knowledge of 2020s open-source culture, I'd be frustrated as hell.

I didn't fully appreciate how important open-source is as a 21st century movement until I read The Cathedral and the Bazaar this year. Hearing the open-source development story of SQLite also blew my mind a little.

I can see why you're so energized by open-source firmware.

Yes! The final two frontiers (or, among the final two frontiers, anyway) of proprietary software are firmware and EDA software. I am (as you noted!) very bullish about both, and I think that open firmware + open EDA will lead to a new golden age of HW/SW co-design -- it's a great time to be a software engineer, and it's only going to get better!

> So you wouldn't be "heading to the recruitment stalls" for any of those companies except potentially Microsoft.

Late '91 I was graduating soon and I did go to the university recruiting stalls of Sun and IBM and Motorola. DEC was a presence but I loved SunOS and didn't like VMS so didn't talk to them. Never considered Microsoft since I hated them so much already back then.

Ended going in a different direction due to a graduate degree scholarship but thinking back maybe should've joined Sun directly back then. I did later end up at Sun a few years later, roughly same timeframe as you.

Now I could understand this as a blue collar worker, but you’re saying you’d prefer to be a software engineer in 91 than 2021? Come on, of pretty much all the professions were the ones who have reaped the most benefits of the past 30 years

Arguably in 1991, you're getting in on the ground floor but it's a mixed bag. Your new grad salary is probably going to be around $40K or so in the US. And dot-bomb is 10 years in the future.

And that list of companies is sort of a mixed bag.

Adobe has mostly done pretty well through the years.

Sun Microsystems did have a very good decade through the dot-com years but then didn't.

Apple was really struggling at that time.

DEC was on the way down and would be bought by HP a few years later.

Microsoft was about to launch Windows NT so that was a pretty good place to hop on but obviously went on to have a long period of stagnation.

Indeed it predates my entrance to the industry by a half-decade but one thing that's notable about this period is the technology stacks being built during it would in large part be heavily de-emphasized later when the web exploded.

If you were at Apple, you'd be working on a platform (Mac OS classic, or Copland, or Newton) that would be thrown away by the end of the decade.

If you were working at DEC, likewise. (VMS, VAX, even Alpha)

Sun is more complicated, as they pivoted better and took longer to die. That would be a good place to be maybe.

1991 is an awkward year since it's about 2-3 years before the HTTP/browser revolution.

One thing though is that to my eye when I look at what these companies were working on then, it all seems more interesting to me now. The actual employment of a programmer (who wasn't stuck in finance or insurance etc. doing COBOL) had the potential to do some stuff that we at least thought Was going to be groundbreaking back then. NewtonOS and Alpha and Copland, CORBA, PowerPC/PREP, OS/2, research projects like Sun's "Self", etc. it was all exciting stuff. Just very little of it went on to be used later.

For me, Sun definitely looks like the most attractive company on that list. (Though Microsoft might well have been a perfectly good job.)

For one thing, you'd have been much more plugged into the coming internet revolution broadly than any of the others. You'd also have been at least connected to the open source world although Sun resisted aspects of it in many ways.

They were also primarily in Silicon Valley unlike the others.

The 90's had all kinds of tech companies starting. Many of them didn't last, but there was a lot of exciting stuff going on.

Early cell phones. PDA's like Palm. Printer market was hot. Businesses were networking their computers like crazy. The PC accessory market was hot. Video games like the Playstation were about to come out. Dial up online services and then ISP's. The web appeared.

I guess if you had hindsight it would be great, just live very frugally, choose the right company and above all buy as much real estate as possible in SV :)

> Your new grad salary is probably going to be around $40K or so in the US.

Of course $40K then is $80K now, and you were working at 9-5 at BigCo, with a pension plan. Interest rates were 3-4x what they are now, the value of the house that you spent a couple of years salary on is probably going to quintuple, and the stock market runup over the next 30 years is going to be unreal.

>the stock market runup over the next 30 years is going to be unreal

Not counting the stock market plunge especially in tech in the dot-bomb era when there's a good change you'll also lose your tech job and very possibly be very underemployed for a few years. Of course if you hold on through that (and 2008), you'll come out well on the other side.

And if you were 20 in 91, you probably don't have a house in ten years in pricey (just not eye-watteringly so) SV 9 years later when the bottom falls out of the market.

You might also be describing 2021-2051.

Well, except pension, which were largely gutted in the private sector by 1991 (it was the plot of the 1987 movie Wall Street).

I moved from Wall St. to the West coast in the 90s. Both paid above average programmer salaries for the time, but S.V. was the place to be for tech. So much going on, so much demand.

I'm sure the day-to-day experience of building software is generally much better in 2021, but as others have noted there are other reasons to start in 1991.

- Particularly interesting point in history for software (pre-Netscape!) - Build lots of experience before Google, Netflix, and FB are even born. - Incredible compounded returns in software stocks (and great returns in housing)

2021 is a great time to be a 20yr old software dev, but it's also a great time to be a 50yr old software dev who has 30 yrs of experience, stock market returns, and housing investment.