It’s weird, I hear there’s rodents involved… I don’t know much more than that. I’ve been interested in things related to package management for a long time. One of my earlier things in my open source world was I was the lead engineer who built the Git hosting platform that’s still used by the Drupal project, so I have this — it’s not technically package management, but it’s quite adjacent; it’s source sharing, and all that. So I have this long history with dealing with communities and them having code and wanting to share it.
A few years ago – I think actually it was around the first GopherCon, 2014… I know Matt Butcher and Matt Farina – all three of us were co-workers back in 2010 or so, and they’re of course the authors of Glide now. At the first GopherCon we were sort of talking, “Hey, we could work on this package management space that really desperately needs it”, although in 2014 — I think it’s fair to say that at that point there was not as much clear public understanding or dialog around it being such a problem.
So they started to work on Glide; I didn’t really come in until later… But there have been all these different projects all over the place from a bunch of different people that tackle at least part of the general problem.
For me it was end of December 2015 that I started getting — I can’t remember what the precipitating event was, but I remember being really annoyed (that much is clear in my mind) with some tool that I was using. It may have been Glide, maybe it was GoDep – it doesn’t matter. But I decided “Okay, I’m gonna try to describe how we should solve this problem”, so I spent six weeks writing this essay which turned into this 13,000-word monstrosity that’s on Medium, that I published February of last year.
I’m proud to say that I stole five years of productivity from the world, because that’s how many hours it takes for the number of people who read the whole thing to read the whole thing. [laughter] It’s my favorite way to think about it.
[00:03:48.29] This is the thing, too… This is a complex social problem as much as it is a complex technical problem, and I think it’s fair to say that there was a lot of acrimony over this problem a year ago. The approach that I decided to take was I was gonna write this essay and then try to crank up this library that is now GPS and is the thing that’s powering the dep tool, and try to use that as a way of — instead of just adding another tool, to use that as a platform for bringing together the different tool maintainers and sort of creating a common shared conversation that let us move in more of a consensus direction instead of this fractious thing that had been going on for so long.
I feel like I had an MVP with GPS around GopherCon last year. When we tried to wrap this up, the shortest summary from there is that Peter Bourgon decided to convene this package management committee, which started meeting September. That was me and Ed Muller and Jess Frazelle and Andrew Gerrand, and we were all on the phone a couple times a week, trying to hammer this thing out.
Around October last year we started actually implementing the dep tool on top of GPS and then we released in early January. Now we’re on the mad dash towards making this thing real and official, that is integrating it into the Go toolchain.