text stringlengths 0 2.35k |
|---|
**Alexey Palazhchenko:** Well, yeah... Mat can't be imitated, for sure. So my unpopular opinion would be about generics in Go... |
**Natalie Pistunovich:** Dun-dun-duuuun...! |
**Alexey Palazhchenko:** Yeah, so quite a lot of people consider generics to be great in Go, and it works great for them... And even more people consider that generics were a bit rushed, and are afraid of them, I would say... So for me, personally, generics went, I would say, sideways. There are a lot of things in Go t... |
To give you an example, let's say I have something like a JSON object and I want to traverse it by using paths and indexes, field names as strings and indexes as integers... And I can't write generic functions that accept a combination of them. It can either accept a list of paths, or a list of indexes. But I can't wri... |
So my opinion would be like generics went in a direction I did not want them to go. So they did not change a lot of how I write Go, and I don't expect that to change anytime soon. |
**Natalie Pistunovich:** Okay... My follow-up question would be do you think your unpopular opinion will be a little bit unpopular, or very unpopular, or actually it may be popular? What is your prediction? |
**Alexey Palazhchenko:** I would say a little bit unpopular. |
**Natalie Pistunovich:** Okay... So you're not going to necessarily make it to the rank of the most unpopular ever opinion... |
**Alexey Palazhchenko:** Who knows...? Maybe generics will be fixed with the help of Copilot in the next release. |
**Natalie Pistunovich:** \[laughs\] Yeah, interesting. I have to say, the last couple of episodes had unpopular opinions on the field of generics; what a surprise... Well, that was an interesting conversation, for sure. If anybody who has listened to this episode and wants to try Copilot but is not there yet, so you ca... |
Yeah, that would be wrapping it up for today, so thank you all for joining us... And join our outro tune. |
• Introduction to static checkers |
• Rubber duck debugging technique and its uses |
• Importance of taking breaks in problem-solving (including going for a walk or engaging in meditation) |
• Matan Peled's background: Ph.D. candidate, research on meta programming and static analysis, experience working in industry and academia |
• Discussion of programming language design and creating animations using a custom language |
• The speaker's past project didn't meet expectations due to lack of usability and tests. |
• The importance of being practical and useful vs. exploring new ideas without immediate utility. |
• Startup culture and its emphasis on flexibility and taking risks. |
• Static analysis, including its purpose and challenges in dynamic languages vs. typed languages. |
• The theoretical basis for static analysis, citing the halting problem and Rice's theorem. |
• Meta-programming using static analysis as a focus of the speaker's research. |
• Static analysis techniques for code |
• Machine learning (ML) applications to code, such as GitHub Copilot |
• Formal methods for static analysis, including type checking |
• Limitations of ML approaches, such as lack of explicit knowledge about code structure and syntax |
• Potential next steps in AI-generated code development, including creation of static and dynamic checkers |
• Discussion of taint analysis as an important tool in recent years |
• Mention of static checking and its use to prevent bugs early on in the development process |
• Examples of types of bugs that can be caught through static checking (e.g. memory allocation issues, multi-threading problems) |
• Overview of static checkers for Go, including Staticcheck and Errorcheck |
• Goal of creating a custom static analysis tool with user-definable rule sets |
• Challenges in performing points-to analysis, including aliasing and dynamic type determination |
• Comparison of static analysis at compile-time vs. runtime debugging |
• Reverse debuggers allow stepping back and going back in time to see what happened before |
• They keep snapshots of operations, but only at certain points (e.g. before input/output) |
• Dynamic analysis involves using information from compile-time plus real-time values |
• Instrumenting is a form of dynamic analysis that involves adding code to track program behavior |
• Structured logging can be seen as a form of dynamic debugging or tracing |
• Go's open-source nature allows for understanding and use of its toolchain packages |
• Self-hosting compilers, where compilers are written in the language they compile |
• Concept of "Reflections on trusting trust" and its implications for compiler security |
• Difficulty of detecting backdoors in compiled code |
• Gödel, Escher, Bach book recommendation for exploring self-referential concepts |
• Unpopular opinion: static analysis doesn't work beyond a certain complexity |
• Software engineer job security due to limitations of AI replacing human tasks |
• Next-level abstraction with AI guiding programming and automation of tasks |
• Research on synthesis and machine learning-based program generation |
• Limitations of current AI capabilities in optimizing code and replacing human programmers |
• No code/no bugs as a future goal |
• Difficulty of writing comprehensive tests and ensuring they don't contradict each other |
• Using static analysis to check for test contradictions and identify potential issues |
• Pure functions and their benefits in languages like Rust compared to Go's ability to have side effects in methods and functions |
• Limitations of static analysis when dealing with input or unknown values |
**Natalie Pistunovich:** Good evening, morning and afternoon, everyone, and welcome to the episode about static checkers. Hello to Mat, my co-host for this episode. How are you doing? |
**Mat Ryer:** Hello to you, Natalie, also. I'm good, thanks. How are you? |
**Natalie Pistunovich:** Great, great! I enjoy all the plants we have in the background. Hello, Matan, our guest for today. Welcome! |
**Matan Peled:** Hello, and thank you. |
**Natalie Pistunovich:** How are you doing? |
**Matan Peled:** I'm doing pretty well. It's kind of late, but... Yeah, it's been a long day. But I'm doing good. |
**Natalie Pistunovich:** And you are joining us from...? |
**Matan Peled:** I'm joining you from \[unintelligible 00:03:11.27\] You can see the sea from that window over there, if it wasn't so dark outside. |
**Mat Ryer:** Oh, wow. |
**Natalie Pistunovich:** So all the way from Israel... And I see you don't have plants in the background, instead you have ducks, which you probably use for duck debugging, as one does. And you have two. One is white, and one is red, for everybody who's listening but not watching... So this is the one who says "Just fo... |
**Mat Ryer:** \[laughs\] |
**Matan Peled:** I have a few of them. One of them is a panda duck. |
**Mat Ryer:** Oh yeah, a panda duck. Yeah, it does look a bit like a devil. |
**Natalie Pistunovich:** And the blue one is just in case you're on the blue team, which is kind of decent code. Not too tested, but not force-push. |
**Matan Peled:** Yeah, it seems to balance out the red one. |
**Mat Ryer:** We had an off-site recently at where I work, and it was in Amsterdam... And the souvenir that we got everyone was a little rubber duck, tailored to each person. |
**Natalie Pistunovich:** \[04:06\] Wow. |
**Mat Ryer:** Yeah. It's a very nice gift... But it's very useful for rubber ducking, like you mentioned, Natalie. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.