text stringlengths 0 1.82k |
|---|
**Ben Johnson:** Right, sure. So it's a project that has been active, although I haven't produced anything for quite some time. I did this visualization of Raft (Raft is a distributed consensus protocol). What that means is if you have a cluster of nodes and they need to agree on some data, basically a transaction log,... |
But yeah, they've tried to make a simpler Paxos, if you want to think of it like that. Like a multi-Paxos. But the problem is when you read the paper, probably several times, if you're not in that world of distributed systems, distributed consensus, a lot of these concepts just don't make any sense. I used to do a lot ... |
**Erik St. Martin:** I love it, because it makes the topics a lot more approachable. For a lot of people who don't work in this space, you say 'distributed consensus' and their eyes immediately roll over, right? This is something that's completely unapproachable, it's only for Ph.D. candidates at MIT, Berkeley, or what... |
**Ben Johnson:** \[52:06\] Yeah, that was my next one I was going to do. As far as seeing things… I feel like we have a lot of fancy words in our industry that typically mean really simple concepts, and I think when you can just see a concept... A lot of times it only takes five minutes to just visualize a concept, and... |
**Brian Ketelsen:** Log-structured merge? |
**Ben Johnson:** Log-structured merge-tree. It sounds like a nightmare, but it's not. It's not overly complex when you get into it. The implementation is, but the actual idea of it is not. |
As far as Kafka, I started working on that, and I've tried a lot of different methods for trying to visualize that. I wanted to try to do more for blog posts where it had visualization within it, but it didn't quite fit. I wanted to move to where I was generating out of video, because whenever you put up any kind of co... |
**Erik St. Martin:** I think part of it, too... If it didn't move faster, you could make it more interactive, which would force people to be slower, like if they were supplying the data that went across, or something like that. But I think that's the hard part with teaching anybody anything - if you move faster, for so... |
**Ben Johnson:** Yeah, I totally agree. |
**Erik St. Martin:** Maybe you could have an automated play thing, where you don't have to click play, it just automatically transitions at a set interval, and you could put a slider for speeding it up or slowing it down or something like that. |
**Ben Johnson:** Yeah, that's true. Also the other part of it is it was just insanely hard to make, because I actually had to implement - because I was doing it all the time; it wasn't a set amount of time - I actually had to build the Raft implementation in JavaScript, that actually runs behind the scenes. I'm trying ... |
**Erik St. Martin:** \[laughs\] I see... I didn't think about that. You actually implemented Raft in JavaScript. |
**Ben Johnson:** Yeah, it was terrible. It was a bad idea \[laughter\]. |
**Erik St. Martin:** Brian is giving you evil eyes from behind the microphone; I know it \[laughter\]. I know he's thinking, "Somebody's going to use this for evil." |
**Brian Ketelsen:** Yeah, it's going to be done. Never release that code, Ben. |
**Ben Johnson:** Oh, no, no. |
**Brian Ketelsen:** Never! |
**Ben Johnson:** Ever. |
**Erik St. Martin:** So I know we're running short on time, I think we might actually be a couple minutes over. But one thing we like to do with each show is Free Software Friday, where all of us will go across the board and list off some open source projects that are currently or have made our lives easier in the past... |
So with that, Brian, do you want to go first? |
**Brian Ketelsen:** \[55:48\] Yeah, I'll kick it off with Minikube; github.com/kubernetes/minikube. K-U-B-E. It's a really fast and easy way to get you going on a Kubernetes cluster on your local laptop, which is awesome if you don't want to run 17 VMs to have a cluster. Very nice, simple, lightweight, easy way to get ... |
**Carlisia Thompson:** I can go next. First, let me mention that library I was talking about, it's upper.io/database. |
**Brian Ketelsen:** Oh, the database library earlier? |
**Carlisia Thompson:** Yeah. The not-ORM library. Before I forget too, I want to mention that Ben did a remote meetup event and he showed how to use BoltDB. That was pretty awesome. |
**Erik St. Martin:** Oh, cool, was that recorded? |
**Carlisia Thompson:** It is, yes. So if you go to remotemeetup.golangbridge.org, you will have the past events, and you'll find his event there. It was very good, I saw that one. |
**Erik St. Martin:** Oh, awesome. |
**Carlisia Thompson:** The free software I want to mention today is called Stow, it's part of the… I don't know how to pronounce this… |
**Erik St. Martin:** GNU. |
**Carlisia Thompson:** GNU, yes. I found this out through looking through Brian's and Erik's Dot files. I don't understand how I've been a programmer this long and I didn't know about this. |
**Erik St. Martin:** It's actually used under the covers for a lot of stuff to manage new versions of libraries. |
**Carlisia Thompson:** I thought it only did symlinks? |
**Erik St. Martin:** It does do symlinks, so it can basically symlink the default version of, say, a dynamic library or static library, to the most recent version of it that may exist somewhere else. But it's used for several things like that. |
**Carlisia Thompson:** Interesting. |
**Brian Ketelsen:** But it's very useful for DOT files. |
**Erik St. Martin:** Yeah. I cheated for DOT files \[laugh\]. |
**Carlisia Thompson:** I see, I see. That's why the description here was a bit off from what I was expecting. But yeah, to manage DOT files… Wow, it's so easy. It's so easy to use! Can't believe I didn't know about it. Not I use it to manage my symlinks. |
**Erik St. Martin:** Yeah, so the cool thing about it is you can basically layer symlinks. The way I set it up in my DOT files is I have a DOT file GitHub repository, and all of my configuration files are stored in this DOT files directory, but then they're managed… The path structures based on the way it would be in m... |
**Brian Ketelsen:** Yeah, it's winning. |
**Carlisia Thompson:** My biggest problem with symlinking manually was to remember how to un-symlink things. With Stow there's a delete command, and it's so easy. |
**Erik St. Martin:** Yeah. To un-Stow it? |
**Carlisia Thompson:** Yes, or un-Stow; I don't know, I don't know remember. I only did it once. I set it up super quickly and it worked, and I never had to look at it again. That's why I thought it was so great. |
**Brian Ketelsen:** Stow/d, I think. |
**Erik St. Martin:** Yes. |
**Brian Ketelsen:** Awesome. |
**Erik St. Martin:** So we have another lover of Stow. |
**Brian Ketelsen:** Yay! |
**Carlisia Thompson:** Yes, thank you. You did a great job explaining how it works, Erik. Thank you. |
**Erik St. Martin:** I set up that readme a long time ago. At the time, I think Brian was wanting to steal some DOT files, and a couple of other people we worked with, so I wrote up a thing because I was having to go to people's desks and help them, so it was just easier that way. |
So Ben, do you have a project you want to thank? |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.