text stringlengths 0 2.35k |
|---|
**Christian Simon:** Yeah, so I have an unpopular one on the Go execution tracer. So while it is very colorful, and I think you can learn a lot how maybe your program is running, I think I've never found the problem, and I always started tracing when I was just desperate. And so my unpopular opinion is instead of start... |
**Mat Ryer:** Hey, that's interesting... Is that unpopular, Kemal? |
**Bartlomiej Płotka:** Not in my book, I guess... |
**Mat Ryer:** Okay. Well, we'll test it, Christian, and we'll see. Kemal, did you bring an unpopular opinion today? |
**Kemal Akkoyun:** I brought a bunch of them. I don't know where to start, but I guess maybe I can just keep it simple and focus on the things that we talked today... |
**Mat Ryer:** It'll keep Bartek happy... |
**Kemal Akkoyun:** \[55:50\] Yeah. This one for Bartek. So I think Go should give certain APIs to control memory for certain use cases, right? Go runtime does a lot of assumptions for us for the memory safety, and it gives us guarantees, and it's really nice with the garbage collection... But sometimes, you really want... |
I know there are some improvements recently. Actually, there are some knobs that you can turn and adjust your garbage collection. On top of that, there is this PR proposal about the Arena support, but still, it would be nice to have more control. I think Bartek is gonna like this, because I know for a fact that he made... |
**Bartlomiej Płotka:** I feel like our opinions were not super-unpopular, because I think I agreed with all of it... And I think there should be some legal agreement we are assigning with you, Mat, that those words cannot be used against us, or something like that. |
**Mat Ryer:** Right... \[laughs\] |
**Bartlomiej Płotka:** But yeah, I would agree with Kemal, Arena is especially good, but I would even like bring a Rust ownership model to like a portion of your Go code. So you could strictly control their scope of living kind of lifecycle of your variables, of your data... And maybe only for 5% of your code, because ... |
**Mat Ryer:** Hm... Yeah, so it kind of make sense then... Like, most people won't have to know about it or even think about it. But when you're doing efficiency work, when you really do want to control it and optimize it, that sort of makes sense. A bit like how you can have Assembly in your Go code, I guess... Becaus... |
**Bartlomiej Płotka:** Yeah. I think we're going in this direction though, so it's good. Arena already is claiming some kind of part of your memory, where it's outside of garbage collection. So it's a precedence. Once that is done, we can go further, deeper into bad things, but also nice for efficiency sometimes, yeah. |
**Mat Ryer:** Well, these all do sound quite popular. Does anyone want to have a go at a properly unpopular one? |
**Bartlomiej Płotka:** No. Before the legal contract, no. |
**Mat Ryer:** \[laughs\] |
**Kemal Akkoyun:** I can go... |
**Mat Ryer:** Kemal, come on. Let's dig deep. Find an unpopular one. |
**Kemal Akkoyun:** Okay. I think the Go team should be an independent third party, instead of something that only Google people actually control. There are a lot of other kinds of stakeholders actually depending on Go. For example, one of the big ones is the CNCF, the Cloud Native Foundation. We work on the open source... |
**Mat Ryer:** Hm... Interesting. I know that that has been a subject before, and I think it is quite a controversial one... But let's have that chat on Twitter. We'll certainly put that clip of you, and then vote, and then that's it; that's really how we do our observability data collecting... |
Excellent. Well, thank you so much. This was great. And we should do this more. You know, we touched on profiling. We really didn't go deep on profiling, and I think that could be also great for a future episode. But thank you so much, Bartek, Christian, Kemal. Absolute pleasure. And thanks for listening. Thanks to our... |
• Charm is an open-source project that aims to make the command line more fun and modernize it for the internet age. |
• The project was founded by Toby Padilla and Christian Rocha, who met while working at BetaWorks, and is now funded by seed capital. |
• Charm has developed a range of tools, including Glow (an app for reading documentation on the command line), Bubble Tea (a TUI framework based on the Elm architecture), and Wish (a framework for building SSH tools). |
• The project's goals include making the command line more visually appealing, modernizing its design and functionality, and making it easier to build new applications and tools. |
• Charm has a set of libraries that focus on frontend development, including Lip Gloss, Bubble Tea, Bubbles, and Harmonica, as well as backend apps like Charm, which is a self-hosted tool for data storage, encryption, and identity management. |
• TUI-based self-serve Git host using Wish and Bubble Tea interfaces |
• Command line learning curve and perception as a barrier to entry |
• Value of investing time in command line skills vs. proprietary technology stacks |
• Business case for targeting developers and enterprises, rather than mass market |
• Comparison to GitHub's modernization of the command line and its success |
• Unique branding and design approach, including 3D backgrounds and "desserty" theme |
• Long-term vision for making production server navigation easier and more accessible |
• Soft Serve: a custom SSH instance on git.charm.sh that allows anyone with a public key to access |
• Wish: a tool for setting up custom SSH servers, divorcing user accounts from machine application |
• Encryption: using symmetric keys and metric encryption keys to ensure end-to-end security |
• Privacy: not collecting analytics or data, prioritizing users' privacy over business metrics |
• Product development: focusing on building products as the product owner or developer, without relying on analytics or A/B testing |
• BetaWorks: a VC firm that provided freedom and support for the team to build experimental projects |
• Funding: raising pre-seed and seed rounds from investors such as Cavalry, Fuel Capital, and BetaWorks. |
• Total funding is 3.6 million |
• The company has a low burn rate due to careful hiring and financial management |
• The team is small, with only six people, but high-impact contributors are prioritized over quantity |
• SSH interface for accessing Git repositories and project tools is being developed |
• Low burn rate allows for long-term innovation and development cycles, avoiding short-term thinking and unnecessary expansion. |
• SSH command line updates for piping output directly into terminal |
• Text user interface (TUI) features and components for building terminal apps |
• Bubble Tea, a tool for building TUIs, and its reusable components |
• Challenges with accessibility, including screen reader compatibility |
• Future plans to address challenges and improve TUI development |
• Inspiration from Kitty terminal's standards and ansi escape codes |
• Challenges of developing for Apple's Terminal |
• Limitations and bugs in Apple's Terminal |
• Importance of innovating the terminal experience |
• Potential for non-text-based UIs using SSH protocol |
• Ideas for mobile apps and rich command line interfaces |
• Use cases for Raspberry Pi and low-powered machines with CLI |
• Vision for a future where SSH apps are composable and easy to build |
• Development of TUIs (Text User Interfaces) with modern programming practices |
• Wishlist project: a tool for managing SSH connections and discovering network hosts |
• Local mode for Wishlist: allowing users to run the app on their local machine and discover their SSH configuration |
• Pi-hole and DNS server management |
• Bubble Tea framework and Bubbles components for building TUIs in Go |
• Advantages of using Go for command-line tooling, including compilation, dependency management, and standard library features |
• Go as the preferred language for building command-line tools |
• Bubble Tea and Bubbles library for building interactive command-line apps |
• Lip Gloss styling and layout library |
• Charm ecosystem for persistence and local storage |
• Integration with SSH for remote accessibility |
• Default servers vs personal servers for storing data |
• Limitations of SSH (e.g. lack of interactive experience, inability to execute code or open files) |
• Possibility of building a Changelog command line app with custom capabilities |
• Use of Wish and Bubble Tea/Glamour for styling and functionality |
• Challenges in making rich user interfaces within the terminal |
• Discussion of alternatives (e.g. Gemini project, custom web pages) |
• The limits of innovation in the command line and how far down to the metal one needs to go |
• Building a terminal is a complex task due to 20 years of backwards-compatibility requirements |
• Exploring new platforms such as VR for terminals, but currently focused on desktop and SSH-related projects |
• Fundamentals of SSH (identity, encryption) and pushing its development forward with Bubble Tea and Bubbles |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.