text stringlengths 0 1.49k |
|---|
These are two pretty wide extremes, but somehow the challenge is "How do we get everyone to participate in the same community and talk about containers together?" It's not always easy. |
**Erik St. Martin:** That's actually a really interesting point I had never thought of. |
**Carlisia Thompson:** Solomon, I wanted to ask you to take us back to the point where you were transitioning from your platform-as-a-service to an open source project... Because what I'm thinking is - and there is no doubt that you were a success story now, which makes it even more fascinating... So what I'm thinking ... |
So you were at this point that people are demanding your service, but they don't wanna pay for it. How was the rationale that instead of saying, "Well, let's change our pricing model" or "Let's make a better offering so people feel that they're getting more value, and therefore want to pay for it." To me seems a super ... |
**Solomon Hykes:** That's a good question. It's not so much that people didn't want to pay for our service and therefore we just abandoned the service... DotCloud was actually a pretty successful product; we had customers, we were growing nicely, and there was definitely a set of customers that were getting value from ... |
Some of them have failed by consolidating with larger companies, and that's worked out great for them, but clearly, there's no giant platform-as-a-service company that's just crushing it and you can point at it and say "I wanna compete with those guys." So there was a general sense that we were in a market that had no ... |
\[12:20\] The problem when you're doing everything for your customer - it's a cookie cutter solution; you have this monolithic platform that does everything for you, and it's take it or leave it. It's super convenient, but if you wanna customize things, you can't; you have to leave or wait for DotCloud to add that cust... |
One comparison I make often is a regular toy and Lego. We had a specific toy that some people liked, but a lot more said "Hey, could I change this or that? Could you let me build my own toy? Just give me the Lego." So we started experimenting with that on the side. We did a side-project - which became Docker - to say "... |
In the end, there was more interest for the new thing than for the old thing, and we just made the call. In your question you asked about the leap of faith - the leap of faith came when we had two things which were both viable, and that's why the decision was hard. It would have been an easier decision actually if DotC... |
**Adam Stacoviak:** I would have thought you would have said that Docker was clearly a bigger play, that's what I thought you would have said. |
**Solomon Hykes:** Well, that is the conclusion that we reached... |
**Adam Stacoviak:** Okay... |
**Solomon Hykes:** ...and that's why we made the leap of faith, but at the time it was not obvious at all. |
**Adam Stacoviak:** It was hard to see that. |
**Solomon Hykes:** We pivoted two weeks after -- you know that lightning talk at PyCon? |
**Adam Stacoviak:** 2013. |
**Solomon Hykes:** Yeah, 2013, Docker was introduced... |
**Adam Stacoviak:** So this is [the very first talk](https://www.youtube.com/watch?v=wW9CAH9nSLs) you've given at all about the future of Linux containers, which is what it's called...? And that was Docker. |
**Solomon Hykes:** Yeah. |
**Adam Stacoviak:** And that blew up. |
**Solomon Hykes:** That became our accidental launch, because we were not planning on that being our launch. There's a funny story there also, but the point is I think within weeks or maybe -- yeah, a few weeks after that we launched, and then we pivoted. So it was all within -- |
**Adam Stacoviak:** It was very quick, because when you were on the Changelog you were still talking about DotCloud... That talk actually -- we'll link it up in the show notes for those listening. I'll drop a link here in the Slack channel. That's actually the talk that at the time Andrew Thorp (who was the co-host on ... |
**Solomon Hykes:** \[15:56\] Yeah, I've got a lot of work to do if I wanna be a YouTube star. |
**Adam Stacoviak:** That's right... We're working on it. |
**Erik St. Martin:** I remember first seeing Docker when it was still DotCloud and Brian and I were playing with it, and it was one of those things that you knew this was gonna be big. It kind of changed everything, and I think I've heard -- it might have been that Changelog episode where I think Solomon was mentioning... |
**Adam Stacoviak:** So where does Go fit into this? This is GoTime after all, right? Where does Go fit into Docker? |
**Carlisia Thompson:** Can I ask a question that would precede the Docker story? Did you use Go at DotCloud? |
**Solomon Hykes:** Excellent question. The answer is no, we did not use Go. We were a Python shop, which is why we were presented at PyCon. |
**Adam Stacoviak:** That makes sense. |
**Solomon Hykes:** So DotCloud was written in Python, and although it ran applications of all types, written in all sorts of languages - which, by the way, was our differentiator - because we used containers, we had a common packaging system and deployment system for applications in all languages, which now seems like ... |
Then at some point, we went through a few iterations... The product wasn't quite right, and then I was being a pain in the ass to everybody because I had really strong opinions on how that thing should be designed, so the key engineer on the early project left, basically, in frustration... We had to kind of start over,... |
That was more instinctive than anything and it was not a popular decision at DotCloud at the time, let's just say that... But basically, it was a few things - first, I just had a gut feeling that Go was awesome and I wanted to play with it, just being honest... And there was a deliberate choice also. |
First, we wanted to optimize for contributions. We wanted this to be a very successful open source project, so we wanted a lot of people to contribute. So we wanted something that was easy to pick up and familiar for as many people as possible, and we didn't want anything too extreme or opinionated. I'm not a big belie... |
The other one is in ops and devops tooling, the biggest problem for a long time has been tribal divisions. You have the Python devops tools, and you have the Ruby devops tools, and you have the Java devops tools. At the time at least, those were the three big tribes. And whatever language you picked for your tool of ch... |
**Erik St. Martin:** Oh, I remember Capistrano. |
**Solomon Hykes:** And Java had their own... It was all duplicated, and we wanted to make a tool that everyone could use, so we wanted a language that would compile to a binary, so that -- you know, the old days of the good old UNIX daemons, SSHD... Who cares what language it's written in? It's a binary; you just drop ... |
All these things rolled into one and we just went for Go all in, and Docker was my first Go project, and obviously, it was a good bet. We surfed the Go adoption wave for sure, and we contributed back. So that's why we picked Go. |
**Adam Stacoviak:** \[20:03\] On that note, Erik, you mentioned you were at a meetup recently, and you had a conversation around this... The conversation around \[unintelligible 00:20:07.18\] that Solomon is pointing to, which is Ruby is probably pretty popular - and Matz even says so - because of Ruby on Rails. Do we ... |
**Erik St. Martin:** So we were basically talking about the adoption curve of Go, and Solomon, you just kind of mentioned you picked Go because you wanted a lot of contributors... But back then, we're talking about 1.0 had only been released maybe within a year of the development, and I really think that 2014 to 2015 i... |
So I really feel that year was like a perfect storm of things. Conferences started popping up, you had Docker, and... |
**Solomon Hykes:** Yeah, definitely when we decided to use Go it was at the very end of 2012. |
**Adam Stacoviak:** That's early. |
**Erik St. Martin:** So a month after 1.0. |
**Solomon Hykes:** And definitely at the time it was not an obvious -- it wasn't hyped; we weren't seeing "Oh, we've gotta get on that. We've gotta jump on that." It was more, "Hey, I'm excited about that. The hacker in me just wants..." Sometimes there's just a tool or a language, you wanna use it, and then after the ... |
**Adam Stacoviak:** We're about six minutes out from our first break - or I guess our only break during this show - but take us back to some of the arguments internally that happened around your choice. To be clear, it was your choice to choose Go versus Python for the future Docker? |
**Solomon Hykes:** Yes. |
**Adam Stacoviak:** What were some of the sales things you had to do to sell the team on this choice, especially since you're pivoting and you're betting so much...? There was a lot of pressure to get it right, so how did you make this happen? |
**Solomon Hykes:** So first, at the end of 2012 we were not pivoting yet, right? To be clear, we were a company of about 20 people, and 18 people were working on DotCloud. Then you had me and one engineer doing this side-project. So internally, for a while, it was like "Solomon's pet project. He wants to keep coding, l... |
**Adam Stacoviak:** "...let him do it." |
**Solomon Hykes:** "...let him do his thing, basically." So when I said, "Hey, let's do it in Go", the biggest \[unintelligible 00:23:11.08\] was that it was new and it seems unnecessary to change just for the sake of picking something new. So there was a little bit of a "Get off my lawn, hipster!" kind of reaction, an... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.