It’s really hard to understand today the level of hype around Java and OOP in the 90s. The fact Netscape changed the name from Livescript to JavaScript may be an indicator. This was also peak Sun, they were really driving the web in this era.
The Java hype was totally unprecedented and probably never repeated. The CEO of a big tech company was on network TV promoting a programming language. I heard stories on NPR in the car. My mother called me to ask me "about this Java thing." Java was everywhere and going to be in everything.
In was accompanied by a huge and successful push into universities to make it the standard didactic programming language. Even MIT switched from Scheme to Java.
I was on the crest of this wave—I actually taught myself Java and Scheme the summer before my freshman year @MIT—so I’m fairly certain that’s not quite right.
The intro CS curriculum stayed on Scheme until it switched to Python something like a decade after the Java hype cycle.
What I believe did change was the intro software engineering lab (6.170) switched from CLU (?) to Java around that time.
I started in 2003, learned C++, then did the majority of my courses in C++ with a class in Java with self-taught Python, C#, and PHP along the way for coursework.
Can someone fill me in, but is there still the derogatory "Java school"? I find that silly because most jobs in programming use some sort of managed memory programming language so teaching everyone in Java makes a lot of sense.
That's also why Apple renamed OpenStep to Cocoa. Java was supposed to be the primary development language for Mac OS X (because Java and Cocoa go great together).
For a while you could write OSX stuff in Java. I did. They also had some pretty okay JNI bindings for stuff like quicktime, so I was able to write a java application that used quicktime to load and display videos. I needed to analyze video streams for my dissertation, so I wrote some custom visualization stuff in java that used the quicktime bindings. Good times.
Well, having a platform agnostic runtime, you could throw arbitrary code at, was potentially a great deal. It was one of the founding concepts for ARPANET, which never took off, since nobody wanted to run foreign code from the network on their machines. For a platform agnostic runtime, OO may have seemed to fit well, by encapsulating data, handlers and logic into interacting objects. (This is close to the original use case for OO as described by Alan Kay.)
Fun fact: Java was also renamed from formerly Oak.
> This was also peak Sun, they were really driving the web in this era.
That's not how I remember it. That's how Sun would have liked to see it but it was Apache on Linux or BSD (or even SGI) that was far more prevalent, at least near me. And I spent a good bit of time in the same building as the local Sun dealership. You could not have paid me to use their warmed over and overpriced hardware. And that really is what I associated both SUN and SGI with: companies wasting money.
But hey, we're in a bubble so party like it's 1999. It's fine if your customers are doing the hype thing, but there is no reason to follow them off the cliff. Someone yesterday asked why Bezos doesn't buy one of the big AI players. That's why.
Well, I was and I know plenty of others that did the same. Slackware was 1993. By 1995 we even had Red Hat.
In 1995 NCSA was running just fine and from December onwards there was Apache. I had the first commercial version of the cam software out (which ran on SGI) and a year later it ran on PCs as well.
You were hobbyists, which was a tiny group compared to ISPs, where Sun’s hardware dominated. IBM and HP were competitors but they were less successful with ISPs. Were you in the US or Europe at the time?
You and your friends were probably hobbyists, hackers or small hosters. Datacenters (remember Exodus?) were full of Sun hardware, racked up & labelled with the ‘hot’ startups of early dot-com.
Sun systems were extremely popular in corporate Silicon Valley in the 1990s. VCs would push the more expensive systems onto their well-funded start-ups. Here's a big check, do what the other well-funded start-ups are doing and buy Sun. More cost effective Apache systems were very widely being used by start-ups on tighter budgets. But even Yahoo for example scaled itself on Apache, as did Geocities.
Can confirm. The dot-com era startups I was involved with all had Oracle DB on Sun hardware. Apache was common. Java was somewhat common if you could deal with its slowness. C++ was common if you could deal with memory problems or needed more speed / efficiency than Java.
The VCs I talked to said it was a business decision. They had money to invest, so startups could afford to buy soft & hard-ware that had gone through a QA cycle or two. The VCs figured the exit strategy for most of their startups would be via acquisition, possibly by another startup so they wanted to have a standard environment to make integrating companies tech stacks easier. Or at least less distracting.
I have this vague memory of Yahoo! execs complaining about Viaweb / Yahoo! Store being written in Lisp and management freaking out that they couldn't hire enough Lisp people fast enough. Or at least that's the story that was going around the valley. (Isn't Paul Graham around here somewhere? Or someone who could point to a canonical reference where he talks about Viaweb getting acquired by Yahoo!?)
