text stringlengths 0 2.35k |
|---|
**Mat Ryer:** Oh yeah, I've seen them. I tried it the other day. I put 1, 2, 3 and then it just repeated 1, 2, 3 loads of times. I was furious. I was absolutely living. \[laughter\] But you're actually right - if it's doing that with the code... I mean, if it's got things like that points to analysis, where it understa... |
**Matan Peled:** Oh, yeah. |
**Mat Ryer:** It could give you an early version, and over time just keep sort of improving it, and things. And that does get very exciting, when you think of that running at scale. |
**Matan Peled:** So again, my unpopular opinion is that that will never happen. That the best it can do is sort of point you in the general direction of saying "Well, maybe you wanna look at this. This could be a good place to look at." But that it will never be able to do that by itself. It will never know enough abou... |
**Natalie Pistunovich:** Interesting. |
**Mat Ryer:** Yeah... Because even like programs we write - they contain bugs. I had a manager once that said he doesn't want any more bugs in code. |
**Matan Peled:** Yeah, so just don't write any more code. No more bugs. Easy. |
**Mat Ryer:** Yeah, there you go. |
**Natalie Pistunovich:** No code. No code is the future. |
**Mat Ryer:** There's definitely truth in that, but genuinely though... Like, yes, it's about whether the program does what we want it to do based on criteria that's external to that program. So in a way it's not available to it to know that. But I don't know, could you write a test...? |
**Matan Peled:** But even the tests don't specify the program completely, right? We all know that writing tests is hard. |
**Mat Ryer:** Writing good tests is hard, isn't it? |
**Matan Peled:** Writing tests that are a good specification is even harder, because - well, if you tell it to multiply, you get two, it goes to four... Then, "Okay, yeah, I can write a program that always outputs four." That works, right? |
**Mat Ryer:** Mm-hm. That's why you need more than one test case. |
**Matan Peled:** Exactly. |
**Mat Ryer:** Yeah. |
**Matan Peled:** It will almost be adversarial, in that it will always find a way that it can do the thing that you don't want it to do, instead of the thing you wanted it to do. |
**Mat Ryer:** Yeah. Also, what you want can change too, over time. It's quite interesting. Could using static analysis -- could you check to see that tests don't contradict themselves? |
**Matan Peled:** Oh, that's an interesting question. I suppose you could, but it depends what you mean by contradict themselves. You could use static analysis to extract them somehow and compare them, and see... Yeah, you could definitely do that. You could see if what they say about the method that they're testing - i... |
So a static analysis -- let's say you have a method that returns some integer... Then we have various sorts of integer analysis that can give bounds. This output is between 0 and 8. Give it like an interval. That's interval analysis. There are more complex types of integer analysis that try to figure out what the value... |
**Mat Ryer:** Yeah, that is interesting. I guess that's why pure functions, like that Rust has - that must be a much easier language to work with than Go, because they can have side effects to the methods and functions in Go. |
**Matan Peled:** So static analysis doesn't really care about the side effects, because it's not executing anything. If you're reading input or something like that, then obviously you have no idea what value it is, or what value it might be... But you just mark it as any, as top, and keep going. Yeah, it can be anythin... |
**Mat Ryer:** Hm. That sounds really cool. |
**Natalie Pistunovich:** Alright, folks, that was very interesting, and that developed even in a more interesting direction after the unpopular opinion. I already wonder what is the next episode we're gonna do about this. Until then, thanks to everyone who joined us. Have a great rest of your day! |
• Intro and discussion about Jerod Santo trying to impersonate Mat Ryer |
• Introduction of guest Ian Lopshire and discussion about his background with Go |
• Kris Brandow shares his experience learning Go and his initial confusion with the := syntax |
• Jerod Santo discusses his own background and experience with various programming languages, including Perl, Ruby, JavaScript, and Elixir |
• Discussion about why Jerod is interested in Go, including its simplicity and suitability for building command line applications |
• General discussion about Go and other programming languages (Rust) |
• Arrays vs slices: arrays are fixed-size groups of data, while slices are resizable |
• Why there are two separate concepts: Go is about simplicity, arrays provide fixed-size properties, and typing in the language makes it complicated to make arrays resizable |
• When to use each: use slices unless you know better or need an array's specific properties |
• Quick assignment syntax: allows for flexibility in type declaration by not requiring explicit type declaration upfront |
• Quick assignment vs explicit assignment and its implications on variable shadowing |
• Use cases for shadowing: within loops, with goroutines, and in error handling |
• Risks and pitfalls of shadowing: bugs, implicit behavior, and debugging challenges |
• Modules as the preferred method for dependency management in Go |
• Current state of dependency management in Go, including history and best practices |
• Module initialization in Go projects |
• Idiomatic Go (normative coding practices) |
• Capitalization conventions in Go (e.g. ID capitalization, constant naming) |
• Use of camel case vs snake case in Go |
• Avoidance of globals in Go code (especially public and library-level globals) |
• Skinny main functions and testing philosophy in Go |
• The Go language has idioms that are nuanced and provide depth to the language. |
• Some idioms in Go may not be effective for every developer or situation. |
• Reading the Go Proverbs is a good way to get started with learning Go idioms. |
• Go is not well-suited for building dynamic web applications like Rails, Django, or Laravel. |
• The strong typing and lack of meta-programming features in Go make it less suitable for complex web development. |
• Go can be a good choice for building custom solutions that require more engineering effort upfront. |
• Focusing on getting started quickly vs. building maintainable software is an important consideration in choosing a language for web development. |
• The benefits and drawbacks of building fast versus architecting a system for long-term use. |
• Why there isn't a direct equivalent to Rails or Django in the Go community, despite its popularity. |
• The idea that existing frameworks and systems can make it less necessary to build new ones. |
• Discussion on the pros and cons of using Go's standard library versus external libraries or frameworks. |
• A lack of certain features or tools in Go, such as better API building tools and database access methods. |
• desire for data to be distributed and accessible without worrying about storage |
• critique of SQL and its underlying model, leading to a desire for alternative solutions |
• interest in using code generation to simplify interaction with data |
• appreciation for languages with built-in features such as DSLs or functional programming facilities |
• discussion on package management systems, including Rust's crates system |
• mention of missing features from other languages, specifically iterators and FP functions |
• debate over handling errors in Go, particularly the use of if err!= nil versus try/catch |
• Discussion on handling errors in code |
• Importance of adding context to error messages |
• Criticism of repetitive if err!=nil {return err;} statements |
• Proposal for a Go fork due to perceived shift from practical to academic focus |
• Naming suggestions for a potential Go fork (e.g. NoGo, Gone, Stay, Og) |
• Twitter polls and their validity |
• Unpopular opinions on social media |
• Voting as a way of self-expression and importance of sharing opinions |
• Show wrap-up and guest thank-yous |
**Jerod Santo:** Hello, and welcome to Go Time. I'm Mat Ryer, and I've spent my entire life working on my mid-American English accent for this very moment when I could imitate Jerod Santo... No, I'm not Mat Ryer! Mat is -- I don't know where. But I am here, I am Jerod, your humble producer, coming out from behind -- wh... |
**Kris Brandow:** The soundboard, maybe...? |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.