CppCast - FluentC++

Episode Date: September 14, 2017

Rob and Jason are joined by Jonathan Boccara to talk about the FluentC++ blog and the benefit of doing daily C++ talks at your office. Jonathan Boccara is a passionate C++ developer working for Murex on a large codebase of financial software. His interests revolve around making code expressive. He regularly blogs on Fluent C++, where he explores how to use the C++ language to write expressive code, make existing code clearer, and also about how to keep your spirits up when facing unclear code. Jonathan loves writing, making videos, reading programming books, hanging out at conferences, meeting people, learning new languages and making trainings and presentations. News C++17 is formally approved Clang 5.0.0 Released (already on the compiler explorer) Two-phase name lookup support comes to MSVC CppCast compiles on godbolt Jonathan Boccara @JoBoccara Links FluentC++ STL Learning Resource Spread Knowledge in Your Company with Your "Daily C++" Murex Going Native 2013: Sean Parent "C++ Seasoning" Sponsors Backtrace JetBrains Hosts @robwirving @lefticus

This episode of CppCast is sponsored by Backtrace, the turnkey debugging platform that helps you spend less time debugging and more time building. Get to the root cause quickly with detailed information at your fingertips. Start your free trial at backtrace.io.cppcast. And by JetBrains, maker of intelligent development tools to simplify your challenging tasks and automate the routine ones. JetBrains is offering a 25% discount for an individual license on the C++ tool of your choice, CLion, ReSharper, C++, or AppCode. Use the coupon code JetBrains for CppCast during checkout at JetBrains.com. Episode 118 of CppCast with guest Jonathan Bocra, recorded September 13th, 2017.

In this episode, we talk about big updates from Klang and MSVC. Then we talk to Jonathan Bulkera. Jonathan talks to us about his C C++ developers by C++ developers. I'm your host, Rob Irving, joined by my co-host, Jason Turner. Jason, how are you doing today? Doing good, Rob. How are you doing? Doing good. I don't talk too much about my day-to-day job, but I've been pretty busy lately. My software, the software I work on for the company, is used by first responders.

So I've been getting a lot of feedback out of Irma and Harvey. Wow. Wow. Yeah, but it's good. We've been making some good fixes and improvements based on all that. Well, it seems like a great opportunity, really, to improve the product. Yeah, kind of crazy because we're trying to put out a release right now as is and just getting all this extra feedback as we're trying to put out the release makes things a bit stressful but i think we're uh we're finally got the release out the door and i hope we're done with hurricanes for a little bit uh well i mean hurricane season still has a month left so we'll see but yeah we'll see or however long it is i

know it goes at least until october You know, I have a random comment, random thought that I had earlier yesterday, I think it was. We have a lot of guests from Europe, and we've had some from New Zealand and Australia, and the US. And I don't think we've ever had any guests from Asia, or from Africarica or from india no and we've had guests from russia so i'm just thinking if we have listeners from a part of the world that we have not yet had on the show reach out to us see if you got something you want to talk about yeah that's if we can figure out the time zones c++ developers all over the place i'm sure we could uh find some from some countries we haven't talked to yet

yeah well at the top of our episode I'd like to read a piece of feedback this week we got an email and I can barely try to pronounce this name it's from Zbigniew Skoran so I apologize for butchering that and he writes in

hello I really like your podcast the episode about Volta was especially interesting since it touched on things we don't normally hear about and yet are highly relevant. Thanks for that. As for my guest proposals, there are a group of well-known programmers that are very anti-C++ and even more so anti-object oriented, and they have very good reasons for that. Even though I don't agree with them fully, I've learned a lot by looking at things from their perspective, and it's improved how I use C++ and

helps me avoid many pitfalls. So we send along the recommendations, and some of these we've heard of. John Blow, a well-known game developer who we may have reached out to him once. I don't think I'm not sure if he responded or if we just weren't able to set that up.

But also Casey Muratori, who invented immediate mode GUIs, and Fabian Gison, who has an interesting blog about high-performance code from graphics to compression to assembly. I think I know him from Twitter. Oh, okay. Well, we'll have to reach out to them. Do you know him as being anti-C++? If that's who I'm thinking of, I never noticed that, but I could just be thinking of the wrong person also.

Well, it could make for an interesting show to talk to someone about why they have that perspective. Yeah, it could be interesting. I will say, just for the record, that it does seem to be that guests who would be anti-C++ don't tend to show much interest in coming on the show. Yeah, yeah. It's worth a shot, though. Yeah. Worth a shot. Well, we'd love to hear your thoughts about the show as well. You can always reach out to us on Facebook, Twitter, or email us at feedback at cpcast.com.

And don't forget to leave us a review on iTunes. Joining us today is Jonathan Bacara. Jonathan is a passionate C++ developer working for Murex on a large code base of financial software. His interests revolve around making code expressive. He regularly blogs on Fluent C++ where he explores how to use the C++ language to

write expressive code, make existing code clearer, and also about how to keep your spirits up when facing unclear code. Jonathan loves writing, making videos, reading programming books, hanging out at conferences, meeting people, learning new languages, and making trainings and presentations. Jonathan, welcome to the show. Hey, it's good to be here. Definitely.

Thanks for joining us. Looking at your bio and you say large code base, and I've noticed a huge range of definitions for that. So I'm curious, what do you consider to be a large code base and i've noticed a huge range of definitions for that so i'm curious consider to be a large code base well our code base is in the tens of millions of lines of code that counts as big yeah yeah well you've got bigger of course but it's fairly big so it's big but not necessarily Google big, I guess. Yeah, exactly. Okay, well, Jonathan, we've got a couple news articles to discuss,

and then we'll start talking to you more about your blog and some of the other stuff you've got working on, okay? Sure. So we actually have a really big news week this week, Jason. Three pretty major announcements. The first one from Herb Sutter's blog is the ink is officially dry on C++ 17.

Well, almost officially dry. Yeah, so the draft standard came back with 100% approval with just editorial comments, which is like spelling and grammar fixes. Right. I'm kind of curious how much work that takes, i i don't think i would want to do it incorporating all the uh editorial fixes

for some reason i just feel like you would get in there and you'd be like oh wait this one editorial fixes actually changes the meaning of this statement or something i don't know well that was the case i think they would just ignore it i would think right i don't know i mean I don't know. Well, if that was the case, I think they would just ignore it, I would think, right? I don't know. I don't think they need to respond to all those. I think they're choosing to. That's my understanding. But it's Richard Smith who has that job.

We've had him on the show before, right? That name sounds familiar. Okay. What I wanted to notice about this post is really exciting news, and I'm really notice about this post is it's really exciting news and really excited about that but I think that most of the work lies ahead for us developers because now we need to learn every bit of it so I think the next question is how do we learn that how do we learn to use it properly

how do we end the pitfalls and everything in it yeah i totally agree this is something i've been spending a lot of time thinking about lately with some of the conference talks i have coming up right how do you use it properly is a big question right i'm sure we'll see plenty of uh cpp con talks on that subject right there's actually only a few talks on C++17 at the major conferences coming up. We might have to wait until the next round and we'll see what happens. Makes sense.

Okay, next big announcement is Clang 5.0 is now available and some of the major new features in here are they implemented the C++ coroutines TS. And even though, as we just said, C++ 17 just became official, they're saying they have full C++ 17 support in here.

Yeah, that's a big deal. Yeah. Yeah, and it ties up well with the previous news article because it's great to have something to experiment, right? Yeah. And also they have officially made the coroutines TS option available, acknowledging that it's a TS in there. Right, right. And, okay, so I don't know how closely you all looked at this, but one of the things under here is clangs diagnostics.

And, uh, that is obviously something that I'm going to look at. And I noticed this very first one dash W cast qual was implemented for C plus plus, but I could not figure out when that warning class is supposed to come up and how to actually get it to warn and C plus plus. Are you guys,

did you spend any time playing with that no i did not play with that no i haven't really so just for the record clang 5.0 is already up on the compiler explorer if you want to go to godbolt.org right that's cool the other thing i saw that was interesting is they're not with this but with version 6, they're going to change the default from what it currently is, which is C++98 to C++14 as the new standard. Which actually makes them the last compiler to make that change for the defaults. Yeah, so they're making this to more closely match GCC's behavior.

Yeah. I'm surprised they're the last ones to make that change. And the next article we have, which is another pretty big announcement, is Visual Studio is getting two-phase name lookup support. And this is something that I think we've talked about a lot recently when talking about their compiler updates. And, yeah, it's finally coming.

It's still a work in progress. But with the 2017 cycle, they're saying they're going to have it fully done. Right, Jason? Yes, but I'm sure there will be... I mean, if you consider what this undertaking is, they've effectively been trying to rewrite the parser for C plus plus for visual studio while still continuing to push out major releases. They didn't do,

like they said, they didn't do a break and then try to do it all at once and then make one major release. So it's a huge undertaking with a lot of potential risk, I think, although probably less risk than trying to just go dark and release a big release all at once. I don't think they could do that because they're such a customer focused company. I don't think they could just disappear

for a year or two working on this. It's great to have two-phase name lookup support in Visual Studio. I'm sure that'll help a lot of libraries that maybe are doing special behavior just for Visual Studio. Hopefully, I'll be able to remove that eventually now. Yeah. And then the last thing, Jason, you pointed this out on Twitter that Matt Godbolt worked on some code on Compiler Explorer that gets the CppCast code to compile.

Yes. That made me very happy when I saw it. Go ahead, Jonathan. Yeah, I was going to say, this one was really funny. When I saw it, at first I didn't get it. I thought, how did he just do that what is this cast i don't know like podcast and then and i saw the trick yeah i had mentioned to him on the side that you know it bugs me to see the c plus plus in our official logo because i'm like it has to be a constant expression.

There has to be a way to get this to compile. And we chatted for just a couple of minutes that it would probably have to come down to some sort of constant expression, plus plus operator overload and whatever. And then 10 minutes later, he comes back with that example. I guess he was bored enough to sort it out. But it's been bugging me for years that I look at that and say,

that can't compile. I apologize that I wasn't thinking about it in that much depth, Jason, when I first just wrote that little snippet. It should have been just podcast CPP, I guess. But to me, that didn't look as good. It's fine. The standard caught up with us it works it works and uh you know other people and other conferences have done similar things and now we know that there are ways to make these things compile yeah well i was very happy about that yeah i'll put the link to that to that in the show notes in case everyone wants to see

how that's done uh under the curtains yeah yeah i saw one comment on twitter that you know it shouldn't return zero it should return like pi or something so if anyone wants to take it to the next level somehow by all means go ahead yeah okay so jonathan um we've talked about your blog posts a lot on this show during the news every week. You managed to be pretty consistent with your blog, posting articles twice a week. How did you decide to post that often? Well, from the beginning, when I started Fluent C++, I was very clear with myself that this wouldn't be a hobby. I wanted it to be a real project to which I would be committed. And its purpose would be for me to learn and also to try and bring something to people.

But the thing is, back when I started, no one knew me. So I had to deliver as much value as I could in a short amount of time. So I set for this rhythm of posting twice a week on Tuesday and Friday. And it seems like that must have been working well for you. It has. Well, you know, about a year ago before the block started i was feeling very proud of myself because i was listening to cpp cast and and today i'm on the show so yeah it has worked but more seriously um i've i have what my purpose was to learn and to read something to people, and I have learned

a ton of things, like, that's bloody amazing. And I've received quite a few pieces of feedback from people that said that some posts had helped them, so I'm really happy with that, and people are giving feedback and proposing things and commenting and sharing, so, yeah yeah it's going okay i'd say i have to imagine also that you're getting like a consistent growth on your readership numbers just uh i mean it seems like that kind of consistency publishing every uh twice a week is huge for getting regulars coming back to your website. Yeah, yeah, you're right. I've seen, yeah, basically constant increase in readership,

except for some posts that went really crazy without really me understanding why. But, yeah, yeah yeah you're right do you have a hard time uh coming up with topics to write two posts every week well i haven't so far and i hope that this will continue um but there's a lot of things that I want to say really. I'm drawing inspiration from several things, actually quite a lot of things. One of them is my daily work so I see problems every day at work, I mean that's a point of work, and I try and solve them most of the time with people and we come up with solutions and those problems I try and extract the generic things that would be useful to other people. been involved in um and also you know when you put things down on paper well not on paper but blog but you know what i mean um you you see holes in your understanding that you see points that you didn't plan and that and and then you understand that your topic that you're tackling is much deeper than you thought initially.

So it happened quite a few times that I would write about something and realize that there would be a big chunk missing and big enough that there would be a new whole post just for treating that. So digging deeper brings content. And I think it's also really interesting because I think people like getting deep into the topics we discuss about on the blog, the technical topics. I think it brings value.

Well, also, I try to give a lot of priority to people's reactions. It happens sometimes that I write something, write an article about something, and then someone would comment on it and say that you could do that differently. And most of the time, it's way better than what I proposed in the first place so when it happens I'm really happy because because it's an opportunity to go further in the topic and that's an opportunity to to develop this idea that the person would suggest and I always try and bring them forward and to give

them credit on the post that i shape after the comment if there's enough material to to create a post from that so i really try to give priority to people's reaction which are invaluable most of the time also um i think a great way to realize things about programming in general is learning other languages. And if possible, languages that are as different as possible from your main language. Like a couple of years ago, I got a bit into Haskell. And this is obviously very different from C++. But it gives you a point of view on programming, which is quite different from what I was used to seeing.

And more often than not, you can draw things from that and apply it to your programming style in your language. So I've learned a lot of things with that. I'm trying to express them through the blog articles. Apart from that, I'm trained with books and other blogs and listening to podcasts. I'm a regular listener to CppCast. I think you guys are doing an amazing job, by the way. And I'm also trying to keep up with watching conferences, CppCon and meeting C++.

There's a ton of content available on YouTube that's just asking to be watched. Yeah. Yeah, I was talking with someone not that long ago who talked about watching all of the videos from a previous CppCon, which is just an outstanding feat in itself i think right so it would have to be i mean there's what six tracks of cpp con i'm just trying to

think how many hours of content that would be yeah and i'm pretty sure there's seven tracks for 2017 if i was looking at the schedule correctly it is a lot of talks that's going to be happening this year. I was listening to how you describe coming up with the ideas for your articles, and you said that as you dig into it, you notice that the problem is perhaps deeper than you originally thought it was. And I started to think it seems almost as though you've made a discipline out of doing, I forget the exact word, is it like rubber duck debugging or whatever? Do you know what I'm referring to?

Like where you just try to just, that's the concept where you like, if you have a problem, you basically describe it to your teddy bear or whatever, just to get your brain talking through the problem. And then that makes you think about it a different way. Right. Yeah, well, I think that one of the most efficient ways of learning something is explaining it to somebody or writing it down, which comes down to the same thing. Because when you do it, the things that you know, before you write them down, they're merely thoughts. And thoughts are deceptive because you may think that you understand something, that you have a global vision on something. And it turns out that this is not true i mean for me it's it's

more often than not it's not true and uh explaining them forces you to structure your thinking and then you start to see it in a different way you take a step back really and then you you got a vision where you see what's missing and what's there. Do you have a favorite article or a series of articles on your blog that you'd like to shine a light on and point to our listeners to? Yeah, sure. Well, I think I talk about a series which is more like a part of the blog really, which I have dedicated to learning the stl

i've called that the stl learning resource i uh i think you've talked about it on the on the show actually uh with stl i mean the guy not the library um well i think that learning the STL is fundamental for writing expressive C++. I came across this idea by watching a talk from Sean Perrin, you know, the C++ Sittling Talk. That was a real eye-opener for me and I guess for a lot of people too. So, I mean, if somebody who's listening now hasn't watched it, then they should rush to their computer and watch it. Yeah, you should try to put a link to that in the show notes, Rob.

That is a great talk. Oh, definitely. In the talk, Sean shows how he transformed a chunk of code that was really hard to grasp at first sight, it down into just one or two calls to functions coming from the algorithms section in the STL, and then the code would just speak for itself.

And that was mind-blowing. And Sean Perrin said that you should know your algorithms. Yes. So I think this is really important. Now, how do you learn your algorithms? I pondered over the question. I tried to learn them myself. And there are quite a few of them.

I reckon that there are about 100 of them before C++ 17. I was going to say, and growing. Yeah, it's growing. And I think just saying I'm going to learn them, I don't think this is the most efficient way to go there. I've thought about it. I've thought about that. And I think that the most efficient way I have found is to learn them by grips

By grips because some algorithm they relate to each other some are slight variations of all the algorithms and When you see them this way by grouping them they will share some commonalities like some common use cases or some common pitfalls that you want to know about so in this STL learning resource I'm putting together my target is to write an article per group of algorithm and and discuss how they are useful how they relate with each other what the pitfalls are and hopefully this makes them easy easier to learn I mean for me it was definitely easier sometimes you you'd

find an algorithm that's so important that I would treat it just by itself, like to transform, for example, because there are so many things to say about that one, which pops up so regularly when you use the STL. So I think that I'd like to, yeah, put forward the STL learning resource as a series. Now, there has been a couple of articles that went crazier than i would think like

how to do during the summer to get better at c++ i think we've talked about that one on the show and this one went mad and there's not one single line of code in it anyway there's also one on meta classes where I tried to summarize the proposal, which I think is really important to know, even though it's not real yet. I mean, it's not available yet. And some posts and smart pointers, I got really good feedback on them and some strong types as well. So that would probably be the one that I would point somebody to. I was curious if you have found the grouping that's on cppreference.com, which I'm not sure if it's the same thing that's also in the C++ standard.

If you're grouping them yourselves or if you're finding value in any of the groupings that exist on other websites or resources. Right. I reckon that the grouping in the standard is essentially mutating algorithms and non-mutating algorithms. And perhaps numeric as well. I think this is about that. And CPP reference has more detailed grouping, I reckon. But I think that we should group them more,

as in in smaller groups, to show how they really relate with each other in smaller groups. Okay. I wanted to interrupt this discussion for just a moment to bring you a word from our sponsors. Backtrace is a debugging platform that improves software quality, reliability, and support by bringing deep introspection and automation throughout the software error lifecycle. Spend less time debugging and reduce your mean time to resolution

by using the first and only platform to combine symbolic debugging, error aggregation, and state analysis. At the time of error, Bactrace jumps into action, capturing detailed dumps of application and environmental state. Bactrace then performs automated analysis on process memory and executable code to classify errors and highlight important signals such as heap corruption, malware, and much more. This data is aggregated and archived in a centralized object store, providing your team a single system to investigate errors across your environments. Join industry leaders like Fastly, Message Systems, and AppNexus that use Backtrace to modernize their debugging infrastructure.

It's free to try, minutes to set up, fully featured with no commitment necessary. Check them out at backtrace.io slash cppcast. Do you have any suggestions for listeners who are interested in blogging and maybe want to get more readers on their blogs or anything like that? Yeah, I do have quite a few, actually. I've been doing that for most of my time over the past few months. So, yeah, I guess I've got a couple of things to share well first of all blogging is a great project um it's it's fantastic because you get to learn

by explaining things as we discussed before and you get to meet all those people i mean don't meet them face to face but you you chat with them online and all those people they're going to bring with them or their point of view on the topic you're passionate about and so that makes you grow really and you're also helping people so I think this is a good experience this is a great great experience now that a couple of things you have to know before starting about being serious about blogging. Actually, I'm going to point out people to great resource before giving the things that I've realized on my own. I've learned quite a few things on the side in the work of John somers which you talked about on the show

when you invited arne mertz i reckon yeah yeah we mentioned john somers and he's the one who has the list of all the podcasts on his blog yeah the one you won yeah i think right um so what he's doing is uh i think this is great and particularly when it comes to blogging. So, people that want to get serious about blogging should go check it out. Now, as far as I'm concerned, I'd say that if you wanted to be serious about blogging, and this will bring you a lot of things, you have to know that this is bloody hard work. It takes time, you know. it takes time to write articles, it takes time to research for topics, to chat with people, respond to comments, take care of emergency that happens

on your website. I mean, it takes a lot of engagement engagement so you have to know that in advance also you have to when you when you write things that you put in front of people you have to know that you will face criticism you will now there are several ways you can react to that and I think that you should really listen to criticism and to what people have to say. Some people they come across as really mean in their comments and I'm sure that most of them don't intend to and that's's okay. I mean, you shouldn't care about how criticism is formulated. You should really discard this aspect and focus on what it is they have to say.

Because most of the time, when I write something, because that's my point of view, when I write something, I'm really close to it and I miss things. And someone who's just discovering my article as a whole, they would see things that I wouldn't see. And they would know things that I don't know. And that's the whole point of it. And so you can really learn from what people have to say. And I think you have to be respectful

to to people that are um criticizing your work and and and try and understand what what it is that they think is wrong and and find a way to grow uh with that then and help people with it um related to the same topic um in such a big project you're investing so much of yourself into, there are moments where you have to push through. Sometimes this is hard. I mean, for me, I'm lucky enough it hasn't been too hard, but that's what I hear about other bloggers in general and some sometimes you just don't feel like don't feel like writing an article you you you want to watch your

episode of game of thrones with your family or something and and you got to push through so this is really important uh you have to see the big picture. It's not just about one article. And if one particular article just doesn't land well, you shouldn't stop everything because of that. You should try and grow from it and push through. that helps you pushing through everything is um if you love what you're doing um i love i love to write really i really do i love to chat with people um and this is something that takes so much of your time that you need to have a strong motivation to to go on with it and the best one you can find is enjoying what you're doing so if you if you're doing for i don't know being famous or making money or whatever i don't think this can isn't this can get you through the whole thing you have to love it and have a strong motivation of learning and and helping people or anything that drives you. You know, I'm aware that it takes a lot of time.

And that's one reason why I was so impressed that you have been consistently doing two articles a week. I'm curious just how much time you spend if you could estimate it per week. It varies from article to article obviously um but in average i i'd say that i spend about two hours and a half writing an article approximately um now that's time for research that that's really varying depending on how familiar with you're with the topic and time to discuss with people, obviously. So at least five to ten hours a week, probably. It's practically a part-time job.

Yeah, yeah. Also, if you want to keep up with the schedule, you should have one in the can, if that makes sense, just in case. So I think you should try and write a bit more than what you publish because if you for some reason one day you can't write then you don't want the blog to stop yeah that's good yeah i've wondered about that uh when we've mentioned a topic on cbp cast and then just like two days later you'll address that question on your blog. And I'm like, well, did he like push his schedule out to fit in the answer? Or what happened there?

Right. I think that was about move iterators. Yes, that's right. Right. Yeah, I pushed the schedule. I put that one in priority. That's funny. I really appreciated that you did take the time to do that.

So what is this concept of daily C++ that you're starting to work on? Actually, I'm sorry, Rob, but there were a couple of things I would add to the blogging thing. Oh, sure. Okay. If I can. Okay. If I can. Okay. There's one thing that is really important. Before that, there's something that you can do to get more traffic because that was your initial question.

I'd say for people that are just starting out, to consider guest posting, which consists in writing posts on a blog that has already a fair amount of traffic I've started to do that on Arne Merz's blog and I've done it on Cipa programmer which is John Sonmez's blog and I think that brought quite a few people I mean some people told me that I knew your blog through that and that, so that definitely brings people around. So I think that's good, and this is also valid for people that don't want to start a blog,

that don't want to get into the hassle of managing a blog, but that have something to get out in front of people, you can just do one-time guest posts on a blog. So I think it's good practice. Actually, if people are interested I'm more than happy to consider guest posting on Fluency++. We've just started to do that this very week actually, and that went pretty well. So I'm happy to talk to people that would have something to say or whether they have a blog or not, really. And one last thing, which is, I think, really important that I want to mention, is that if you have a family and you want to take on such a project

as having a regular blog then you should let them know in advance where you're going because you're gonna spend time you're gonna spend evening typing away at your computer you're gonna spend Sunday afternoons working so you have to be clear with them and they have to agree with that and give you their support you're going to need it one way or the other actually I want to grab this opportunity to thank my wife Elizabeth for that

she's been very supportive and she seems to have an infinite amount of patience with my projects and she's brought on a lot of support. So I really want to thank her for that. Excellent. Yeah, it's definitely very important. So what is a daily C++? Yeah, right. So a daily, first of all, is a new format of presentation that consists in doing short talks of about 10 minutes, but every day. And you give them to a particular team in your company

and you give them right in the office spaces where people actually work so you pick a topic say C++ for example and that would be the daily C++ and every day you do your 10 minutes talk to a company to a team in your company right in their office we've been practicing that for a bit more than a year in my company at Murex and this has become a common practice now. The thing is it's interesting because 10 minutes is a short amount of time and that fits in pretty much everyone's schedule that you can fit 10 minutes in your day to learn things and it's also short enough so that you don't get bored didn't have the time to really and it's

it's designed so that it's super easy to attend. Because when your team is hosting a daily, say the daily C++, for example, then all you have to do is turn back your chair, listen to the daily for 10 minutes, and then carry on with your day's work. And you don't need to spend a second. Now, when you're the presenter of a daily session, then you move on from team to team Now, when you're the presenter of a daily session,

then you move on from team to team every month with the same content so that you give your talks to everyone who's interested in your company. And when you've gone to everybody, then you can start over with new contents um what's um so these are smaller like intimate presentations with just like a few co-workers at a time yeah that's the idea now you could the the days the rooms and times are public so everyone can join and we have typically about 20 to 30 people attending a daily in general which is more than just a team you know so people are joining in what's interesting when you do a daily when you present a daily is that you learn by explaining which is what we've talked about, and you get to meet people, and you meet them face-to-face. Those people are a very enriching experience.

Meeting them is a great experience because they're going to come up to you with questions, and they're going to show you their point of view, and you're going to see a new vision on a topic that you're passionate about. Also, you're going to know who's passionate about the same topic as you. So for me, it's C++. And I'm doing the daily C++ in my company. And by doing that, I've met some wonderful people that I wouldn't have known if I didn't do that. It's the people that would come up to me and ask questions and would chat about C++ and this makes you really

C++ friends, if I may say. So it's a great experience to do a daily as a presenter. Now it takes a bit of work, obviously, because you have to come up with contents. A module for a month is say from 10 to 20 talks so you have to produce that. Well it's a great exercise to produce that. Now not everyone has the time to come up with contents and I I don't know, write about them and give a transcript or something. So I think this is such a great experiment, and I think that everyone should have a go at that. And since I don't want this to hold you back to not have contents, I'm encouraging people to use the contents on Fluent C++ to fuel their daily

sessions. There are actually quite a few posts that I craft with that in mind because I use them in my daily C++. So quite a few of them focus on a precise topic and they fit in about 10 minutes and they're structured in a progressive way. So if you want to make it even more useful to people, a good way I found is to give out the content that you're actually talking about in the form of a transcript, for example, and to compile it into a handout, a booklet. So if people want to use the content on Fluent C++ to their daily C++, then I'm more than happy to compile the post that they're using into a PDF that would be formatted for that. They can print it out and give it to

people. And I think it's something you want to try because you don't have anything to lose. I mean, if for some reason it doesn't work out, then it would just have been an experiment. But if it works out, then you have so much to gain. I really like this idea. You know, we talk about C++ user groups a lot on the show, but it seems that there's so many C++ programmers out there who aren't attending user groups, who aren't going to conferences, and being able to introduce them to a little bit of community

while they're at the office and only takes 10 minutes seems like a really fantastic idea. Yeah, it's gone quite well. Actually, there are other people in my company that picked it up for their own domain of expertise we had a daily Java daily functional programming a daily sequel daily UI and we even had a daily Bitcoin even though we don't do Bitcoin but it's just interesting so we had teams that were willing to to learn 10 minutes of

Bitcoin every day for a month so it just takes somebody who's willing to give the talks and a team that's willing to host it to create a daily session do you find that your company is cooperative with this and they give you the time and flexibility to present oh absolutely this is something that's great with the place i'm working at Murex we really put a strong focus on learning and getting better at what we're doing so when I decided to try this experiment I just came up to my manager and asked him if he was okay and he just told me to try it out. I actually want to thank him for that.

He's actually a fantastic manager, Patrice. And so I tried it out, and people liked it, and we just carried on, and it got bigger, and that just happened. And no one tried to get in our way because this was different. So, yeah, it was great to that extent. Excellent. Do you want to talk a little bit more about what you're doing at Murex with C++?

Sure. So we're working on a large code base, which is in the tens of millions of lines of code. It's a financial software. I'm with the team who's in charge with everything related to equity, so stocks or shares in a company. So we do stock options and futures and stocks and stuff like that, which is a very rich domain and quite interesting and stimulating. On the technical side, the challenge is to make this software evolve. The company has been around for like 30 years. And I don't think there are bits from the beginning that are remaining today in the code, but you've got bits dating from various ages so you have to to make that work to maintain it and and include functionalities as clients are asking

and I've after after the years I found that the single most important thing that would help me in my job would be to try and write expressive codes and This is what got me started on this topic and and this came to be the topic of my blog because if you if you Have a hard time understanding codes if it takes long to understand what you're reading then you're gonna need time to figure out and time is money in a company so you expressive code makes you money

but in a less practical point of view if you if you if the code speaks for itself then it's going to show you where it's wrong, and you're going to spot bugs more easily. So being expressive makes it easier to have a more correct piece of software. And also, and this is a big debate that I see coming up every now and then on the internet or everywhere is about performance. So should you favor a piece of code that's more expressive or that does less? Say, for example, if you have, I don't know, a collection that you would do two things on, is it better to do the two things at the same time when you

run it down or should you run it down two times? That's just one example. And I think that you should always go for the cleanest option. I found this is what works on the long run. I always do that and I treat performance problems afterwards when they come up. When we identify where the bottlenecks are then we fix them and sometimes then yeah we have to make sacrifices to have a application that runs fast. But you don't want to damage your code beforehand, before you know where it matters. So working at Murex has taught me that you really want to favor expressiveness in your code.

So working on this 10 million, multi-million lines of code project that you said has evolved over so uh so many years do you get the opportunity to use uh more modern c++ 14 or even c++ 17 in the code base right well we're actually in the process of migrating our compilers and we're nearly there we should have it anytime soon uh to++11, I reckon, or 14. But it's been a bumpy road. Well, there are lots of codes, and we're sporting several platforms.

So we've got a bunch of compilers to work with. And it's been a huge amount of work to migrate in that direction. But this is something we're really adamant to achieve so yeah we're on our way and nearly there now there is a point i want to make about that about modern c++ i think there's a bias on the internet or in conferences because when you go see blogs or you go watch talks modern C++ is all over the place and when you go back and look at your code if you don't have it then you feel like crap you feel like you're the only person in the world that doesn't have it now I don't think this is true I I meet I meet people you know I hang out in user groups often and I chat with people and

and quite a few of them they're stuck with C++ 03 and they're not even in a process of moving you know for various reasons and also have feedback from people on the blog on freemad c++ that would ask me to produce content specific for c++03 because they that's what they have so i want to say that that there's a difference between what you see on the internet, which is the most modern thing, which are really exciting, and that's what you want to talk about, and what some people have in their production code. So for the people that have C++03, then you're definitely not the only one in the world. And in that situation,

I think that whatever your version of C++ in your production code, I think you shouldn't refrain from innovating in your code. There are so many things to learn about C++03. I mean, there are, I don't know how many books, I've written a post about books to learn C++, and there are so many, so many things to know. So you've got things to learn. And also, you've got things to learn about how to design code. And quite a few things are independent from one particular feature in the language, like how to make an interface clear, for example. So you can still do things,

even if you're not there yet. However, don't get me wrong, I think you should do whatever is within your reach to have modern C++ in your production code. But if you don't have it, don't suffer and don't stop from trying and do great code. Also, about that, I think that whatever version you're working with, you should know and care about what's coming in the language, in the version that you don't have yet. I mean, I don't think that many people have

C++ 17 working in their production code. I mean, Clang just came out with those features, so pretty much no one has it, I suppose, in code but anyway I think that we should all learn it now because it shows us where the language is going and this knowledge we can factor it in our codes even if we're in C++, I don't know, 03 or 11, we can use this. I think that even meta classes are worth knowing about, even if they're merely a proposal that will probably evolve over the years. I think we should know what that is. And this is why I've talked about it on my blog, because it gives you a perspective on the language. And you see how the features that are there today, how they fit in this picture. And this gives you more hindsight about the language. And you can use this knowledge in your code.

Now, even if you don't have the latest version of C++, there are ways to get the features, at least some of them. Some features are extremely easy to get, like algorithms in the STL. You know that some algorithms appeared in C++11? I don't know, like Olive or any of which are so useful? Yes. These are, just take a second to implement. You can just copy and paste the implementation that's

up on cppreference.com into your code and start using them. So this is the most easiest thing you can get. Or just call boost. Yeah, that's what I was going to say, exactly. So boost has features that are coming up in the language, like optional, for example.

It's been in Boost since forever. And it's been in the standard since C++ 17. So you can start using optional if you just take in Boost. And that's what we're doing. You can emulate quite a few feature, even unique para. You can emulate it, not just the same, because you don't have move semantics, but you can do something that's much better than auto PTR.

Heck, you can even implement some ranges, components, that we don't even know when they get into the standard you can you can implement like range adapters you know like filter and transform and that that those amazing thing that you can plug onto a collection and you can like wrap them within each other and have a code that that's more expressive than one you could imagine. You couldn't get that in C++, even C++03, let alone C++11. So I think that if you want to use modern feature and you should have a go for it,

then you have options. Yeah, absolutely. It's been great having you on the show today and talking more about your blog i saw that you're going to be speaking at meeting c++ this year before we let you go could you give us a plug for that talk yeah sure so um at meeting c++ i'll be talking about strong types so by type, I mean using a type that has a specific name that's more meaningful than a primitive type. Say, for example, you're representing a serial number and you're using a string for that,

then you'd use a class, serial number, that does pretty much the same thing as a string, but that's called serial number. There's a lot of things to say about strong types. So I'm going to talk about what they are, that does pretty much the same thing as a string, but that's called to your number. There's a lot of things to say about strong types, so I'm going to talk about what they are, how they can be useful, and also really practical C++ related issues with strong types. How to pass them as reference, how to use them on generic types, how to use them on LOMDATs, for example, how to do conversion, how to use the functionalities of the underlink types,

how they would impact your performance, and that kind of things that would be really practical for strong typing in C++. Sounds great. Yeah, sounds really good. And of course, people can find your blog at fluentcpp.com, right?

Right. Okay. Well, thank you so much for your time today, Jonathan. Thank you so much. Yeah, thanks for joining us. Thanks so much for listening in as we chat about C++. I'd love to hear what you think of the podcast. Please let me know if we're discussing the stuff you're interested in.

Or if you have a suggestion for a topic, I'd love to hear about that too. You can email all your thoughts to feedback at cppcast.com. I'd also appreciate if you like CPP cast on Facebook and follow CPP cast on Twitter. You can also follow me at Rob W Irving and Jason at left kiss on Twitter. And of course you can find all that info and the show notes on the podcast website at cppcast.com. Theme music for this episode is provided by podcastthemes.com.