Two years of Linux Day Milano
Since 2017 I’ve been studying computer science in Università degli Studi di Milano-Bicocca, and after a couple months I joined UnixMiB, the students’ Linux User Group.
I’ve helped these awesome people organize many events, including Milan’s Linux Day, in both 2018 and 2019. They’ve both been great events, especially the latter, we’ve had great fun and managed to attract a lot of people on both occasions.
Linux Day 2018
In 2018, I held my first public conference, titled The Orange Line: in roughly one hour I tried to explain what IBM mainframes’ are and what is their role in the internet’s ecosystem, with a particular accent on the IBM LinuxOne initiative: by outlining their differences in architecture, I tried to uncover the benefits in using a mainframe vs. just a bunch of servers.
The presentation was a mild success, and I’ve got quite a bit of questions. Sadly, there is no footage of the event.
Linux Day 2019
In 2019 instead, I brought a talk about the history of AI titled Weaving a Story, which sought to encompass the whole evolution of this subject since it’s birth in the 1950’s to the adoption of NPUs in smartphones, by outlining the evolution of the hardware that supported each step in this field’s evolution.
I fell a bit short on my premises though: due to the hard limit of one hour I couldn’t talk about interesting technologies like the large vector processors (Cray-1 and such) and why they were NOT initially employed in AI, and most importantly about the architecture of the Connection Machine CM-1, one of the first and few one-bit MIMD manycore architectures, which predated modern trends. I’ve also had to cut a bit short on the last part of the talk, sadly: my fault for not asking for an extra 30 minutes :)
The Importance of History
I have a strong belief that history is what brought us to the point where we stand today: I often feel like history, especially in computer science, is disregarded. We can’t simply cobble stuff together and call it the next big thing, as someone probably already did something similar (and if you’ve not heard of it, probably it’s not that great).
Finding your fil rouge
When telling a story, no matter how complicated it can be, you need a path to follow if you want everything to fall together: looking back, I’ve actually spent more time adjusting my path than finding evidence to support my claims.
A complicated journey
To reconstruct history you need sources you can trust: often a popular computer’s capabilities are overrated by manufacturers and aficionados, while other, lesser known machines get all the dirt, while in some cases they were technically superior. While human history is told by winners, though, computers are designed by engineers. Don’t look for brochures and commercials, but for schematics and insights: the former are only useful to make a point, really. Places like BitSavers and Archive.org are your friends: take your time to read these cryptic Principles of Operations books and Reference Manuals and you’ll be rewarded with a deeper (and unbiased) knowledge on the subject.
Wrapping up
Sorry for the unusual article: I promise I’m gonna go back to talking of boring and uninspiring things soon. In the meantime, if you want to throw away an hour of your life to watch me talk about the history of AI, here’s my intervention (it’s in Italian though, I’m planning to subtitle it).