text stringlengths 0 2.35k |
|---|
**Björn Rabenstein:** But the weird thing is there are actually 98 things in the book, and mine is the 98th. So perhaps I don't really belong to it. |
**Mat Ryer:** Is one of them a section on off-by-one errors? |
**Björn Rabenstein:** Yeah. |
**Johnny Boursiquot:** And the funny thing is, it's not one per author. There's fewer than 97. |
**Mat Ryer:** Oh, that'll save you effort. |
**Johnny Boursiquot:** \[48:10\] There's 98 in this case. Yeah. And then some of us wrote like two or three. \[laughs\] |
**Mat Ryer:** I'll save the flies |
**Johnny Boursiquot:** Oh, yeah, yeah. That's funny. |
**Mat Ryer:** Oh, it sounds good. Well, I recommend people get that book. By the way, I've read it, and yeah, I genuinely do recommend it. Guess what? It's that time... It's time for Unpopular Opinions. |
**Jingle:** \[48:28\] to \[48:43\] |
**Mat Ryer:** Who has an unpopular opinion for us? |
**Björn Rabenstein:** I have a very pressing unpopular opinion. It's because we already use that word that I don't like, and I hope I didn't use it, that's the O word, which is the word observability, right? I mean, my unpopular opinion is that I think this word, if it ever had a reason to be used in our profession her... |
I can very well understand why people came up with this word, because -- I mean, I already talked about this paradigm shift that people sense when this word developers thought monitoring was just an ops concern, and ops people were monitoring by just probing the HTTP ports or something, or pinging the server, or someth... |
So I totally understand that some people felt this urge, "We need a new word for that", right? But then -- I don't know... Observability, arguably, is coming from control theory, and it's -- I mean, I think control theory is super-relevant for monitoring, but it's a rather specific theory, and just taking a single tech... |
I'm more a fan of, instead of overgeneralizing a very specific term, just take the existing terms and appreciate and acknowledge that we have all very different ideas of what that term actually means, and try to find like a wider, more common understanding of that term, right? Now, of course, the problem was also that ... |
So for me, observability is like, if you want to use the term, it's something like a property of your system you want to have, but for me, it's a subset of monitoring, while most people think monitoring is something that might be part of observability, or not even. I don't know. What's his name? I always blank out on n... |
**Bartlomiej Płotka:** \[51:59\] No worries. But I just realized, Bjorn, that I was always saying that monitoring is a subset of observability, and now we are saying it's opposite. I guess I was close. |
**Björn Rabenstein:** Yeah... Cory Quinn. Cory Quinn said-- he always says on his podcast, "Observability is the hipster word for monitoring." And I kind of like that, but I would say observability is the marketing word for monitoring, right? And if engineers call something a marketing word, it's probably doesn't mean ... |
**Johnny Boursiquot:** Do you use the words serverless? |
**Björn Rabenstein:** I do use the word serverless, but it-- |
**Johnny Boursiquot:** So, yeah... \[laughs\] |
**Björn Rabenstein:** Bartek and I just discussed yesterday -- or was it today? Yeah. No, it was today. It's a long day. We discussed how there are serverless applications that are not really serverless, and then there are the true serverless applications that nobody really caters for, because it's so rare that you hav... |
**Mat Ryer:** Well, we'll be tweeting out that opinion to find out, once and for all, if that is indeed unpopular or not. Bartek, have you got an unporpular opinion? |
**Bartlomiej Płotka:** Oh, yeah. Do we have a time for my unporpular opinions? |
**Mat Ryer:** We do. It depends. How long is it? |
**Bartlomiej Płotka:** It depends, yeah. Let's see. Alright, so, quickly... |
**Mat Ryer:** Let's measure it. |
**Bartlomiej Płotka:** So I was thinking a lot -- I have many am power opinions, but some of it that I would truly want to, I don't know, inspire others is that... Stay with me on this. It might sound childish and unprofessional for the start, but... I think the Go language is the best foundation and platform to progra... |
Some of it is that, for example, like - yeah, JSON as a language for configuration. Maybe it's good, but I don't have hints when I kind of point my fields of this JSON on JSON nothing kind of hints me what the structure actually implements. And Golang already had that, right? So why not reusing this and building on the... |
We used configuration in my past job to build kind of configuration for infrastructure in Go, and it wasn't perfect, because there are some features of the language that maybe are too much. You have for loops, you have error checking, and you have a little bit of boilerplate, but if you refine the language, maybe speci... |
And maybe the last example, embedded systems I mentioned. Of course, garbage collection is not very efficient here, but there are already implementations of manual allocations. I think Dgraph created a special allocator. And there's even Vlang, which also kind of removed GC, but kept some of the features that Golang di... |
\[56:14\] So we might have maybe a version of Go that has Rust-like memory ownership, and that'll be amazing, because we keep other features instead of kind of living with totally different decisions that the Rust community made for other stuff, that maybe I'm happy with. So that's my unporpular opinion here. |
**Mat Ryer:** Well, I'll genuinely be interested to see if that is unpopular with our audience, because maybe a lot of people would agree with you. I don't know. |
**Björn Rabenstein:** The Go fans should all agree, right? |
**Mat Ryer:** Go fans. |
**Björn Rabenstein:** But I disagree, for one. I think it's a truly unpopular opinion, because I disagree with it. \[laughs\] |
**Bartlomiej Płotka:** Great. |
**Mat Ryer:** Fair enough. Okay. I'm afraid that's all the time we have for today, but this has really flown, and that's how you know we've had a good conversation... So thank you so much for that. We heard all about monitoring there, and observerbility... The myth of Prometheus - very interesting, the origins there of... |
**Bartlomiej Płotka:** I would love to. Yeah. |
**Mat Ryer:** We'd love to have you. Björn Rabenstein, also, will you come back? |
**Björn Rabenstein:** Yes, sure. It's one of my favorite podcasts. And if I'm in there myself, even better. |
**Mat Ryer:** Is it? |
**Björn Rabenstein:** It's boring for me then, but it's a great honor... |
**Mat Ryer:** Yeah. If you remember everything you've said, that's boring. So if you can try and forget it... Or just listen to it in a couple of years. But yeah, if you just said that you'll be back, that'd be great. |
**Björn Rabenstein:** Yeah, I'll be back. |
**Mat Ryer:** Thank you. |
**Björn Rabenstein:** I should say it with the Terminator voice. Yeah, "I'll be back." |
'**Mat Ryer:** Yeah, that's just your normal voice... Isn't it, Björn? |
**Björn Rabenstein:** Sometimes, yeah. |
**Mat Ryer:** And of course, Johnny Boursiquot. Thanks, Johnny. It's always good hanging out. Björn, good luck with finding John Connor. We'll see you next time on Go Time. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.