WEBVTT 1 00:00:04.840 --> 00:00:07.790 >> Hello. 2 00:00:07.800 --> 00:00:10.590 Welcome to the Haskell weekly podcast. I'm your 3 00:00:10.590 --> 00:00:13.530 host, Cameron Gera, an engineer at ITProTV. 4 00:00:13.540 --> 00:00:16.240 And with me today is Taylor Fausak, 5 00:00:16.250 --> 00:00:19.060 lead engineer. As well as a guest. How you doing, Taylor? 6 00:00:19.740 --> 00:00:22.320 >> I'm doing well, Cam. Thanks for having me on the show this week. 7 00:00:22.330 --> 00:00:25.310 >> So we have a guest. I would 8 00:00:25.310 --> 00:00:26.850 love to have you introduce him. 9 00:00:27.740 --> 00:00:30.650 >> Yeah. I'm very excited to introduce our first ever 10 00:00:30.660 --> 00:00:33.580 guest on the Haskell weekly podcast. We're 11 00:00:33.580 --> 00:00:36.310 thrilled to have him on. Thank you for joining us 12 00:00:36.310 --> 00:00:39.280 today. Matt Parsons, author of Production Haskell. 13 00:00:39.290 --> 00:00:40.210 Thanks for joining us 14 00:00:40.220 --> 00:00:42.950 >> Hi Yeah. Thanks for having me on. Excited to talk. 15 00:00:43.540 --> 00:00:46.480 >> So welcome. You're calling us from Sunny Denver and 16 00:00:46.480 --> 00:00:49.300 we're talking to you from Florida. So we got, like, most of the 17 00:00:49.300 --> 00:00:50.950 United States covered here feels good. 18 00:00:51.440 --> 00:00:53.060 >> The important parts anyway, 19 00:00:54.080 --> 00:00:55.070 >> Very true. 20 00:00:55.080 --> 00:00:57.520 >> I don't know. West Coast listeners 21 00:00:57.920 --> 00:00:58.600 disregard that 22 00:00:58.610 --> 00:01:01.060 >> We love all people. 23 00:01:01.440 --> 00:01:01.760 >> Yeah, 24 00:01:02.140 --> 00:01:03.160 >> we're inclusive here. 25 00:01:03.740 --> 00:01:06.450 >> Um, but yeah, we are really excited to have our first guest on 26 00:01:06.450 --> 00:01:09.240 and particularly talking about this book. But before we launch into 27 00:01:09.240 --> 00:01:12.190 that Matt, I feel like you and I share kind 28 00:01:12.190 --> 00:01:14.890 of a common history or maybe I'm mistaken and just projecting a little bit. 29 00:01:14.890 --> 00:01:17.860 But, um, you have some ruby in your background too, right? 30 00:01:18.240 --> 00:01:19.110 >> Yeah, that's right. 31 00:01:19.110 --> 00:01:20.550 >> The programming language, not the gem. 32 00:01:20.560 --> 00:01:23.410 >> My first real 33 00:01:23.410 --> 00:01:26.340 internship was with a startup in Athens, Georgia, 34 00:01:26.350 --> 00:01:29.290 doing ruby on rails. And that's kind of 35 00:01:29.290 --> 00:01:31.910 where I got my experience building things out with. 36 00:01:31.920 --> 00:01:34.720 I learned how to build Web apps from 37 00:01:34.720 --> 00:01:35.660 the 38 00:01:36.640 --> 00:01:39.360 Michael Hartl's Rails tutorial, 39 00:01:39.740 --> 00:01:42.510 and that is kind of poisoned How I look at all Web programming 40 00:01:42.510 --> 00:01:45.440 ever since then. So now whenever I 41 00:01:45.440 --> 00:01:48.390 look at evaluating something, it's like, All right, I'm gonna build a Twitter clone. We'll see 42 00:01:48.390 --> 00:01:50.260 how easy this is. 43 00:01:50.540 --> 00:01:53.450 >> And that's something that rails make super easy and you say poisoned. 44 00:01:53.450 --> 00:01:56.310 But I feel like it's probably like Ruby and 45 00:01:56.310 --> 00:01:59.100 Rails puts such an emphasis on the developer 46 00:01:59.100 --> 00:02:01.990 experience making that as nice as possible. And I feel 47 00:02:01.990 --> 00:02:04.700 like that's something that Haskell has some 48 00:02:04.710 --> 00:02:07.500 room to improve. Perhaps so pulling some of that is a good 49 00:02:07.500 --> 00:02:08.470 thing 50 00:02:08.480 --> 00:02:11.370 >> for sure, I think I picked up the phrase 51 00:02:11.370 --> 00:02:14.150 poisoned actually from went to rails conf, 52 00:02:14.150 --> 00:02:16.980 2015, with the company, and 53 00:02:16.990 --> 00:02:19.800 we got to see Sandi Metz talk Sandi Metz, 54 00:02:19.800 --> 00:02:22.730 said that she was poisoned by small talk and 55 00:02:22.730 --> 00:02:24.750 was trying to poison Ruby with it, too. 56 00:02:25.540 --> 00:02:28.500 And as far as I can tell, that's like a universally 57 00:02:28.500 --> 00:02:31.270 good thing to do is introduce more small talk idioms into your programming 58 00:02:31.270 --> 00:02:31.860 language. 59 00:02:32.340 --> 00:02:34.640 >> Yeah, I think Ruby was pretty 60 00:02:34.640 --> 00:02:37.610 strongly, um, inspired or maybe stole 61 00:02:37.610 --> 00:02:40.600 from small talk quite a bit. So it's seems 62 00:02:40.600 --> 00:02:41.350 to work out. 63 00:02:41.360 --> 00:02:44.030 >> From what I can tell, Ruby is this weird 64 00:02:44.040 --> 00:02:46.820 combination of Perl and PHP and small 65 00:02:46.820 --> 00:02:49.670 talk and all of the rubyists that really care about programming languages 66 00:02:49.850 --> 00:02:52.350 really want you to forget that it's not just small talk. 67 00:02:54.240 --> 00:02:57.000 >> They say Don't look under the covers, you 68 00:02:57.000 --> 00:02:59.920 know, PHP. You know, that kind of stuff. Leave it alone. Let's 69 00:02:59.920 --> 00:03:01.260 just focus on small talk. 70 00:03:01.740 --> 00:03:04.640 Well, my background's in javascript, 71 00:03:04.640 --> 00:03:07.250 So you know, web app 72 00:03:07.260 --> 00:03:08.000 programming? 73 00:03:08.870 --> 00:03:10.200 >> sweet. 74 00:03:10.210 --> 00:03:12.900 >> But now we're in Haskell. So 75 00:03:12.900 --> 00:03:14.790 that's that's the good thing. 76 00:03:16.040 --> 00:03:18.710 >> Yes. So, uh, since we had this kind of 77 00:03:18.710 --> 00:03:21.490 shared background, I'm curious. What drew 78 00:03:21.490 --> 00:03:23.160 you to Haskell in the first place? 79 00:03:23.940 --> 00:03:26.690 >> So when I first started learning how to 80 00:03:26.690 --> 00:03:29.480 program, I was doing java at the university 81 00:03:29.490 --> 00:03:32.110 and I decided that there were Well, I didn't 82 00:03:32.110 --> 00:03:35.000 decide. I found out that there were no java internships 83 00:03:35.010 --> 00:03:37.780 in Athens, Georgia, where I was going to school 84 00:03:37.780 --> 00:03:40.340 at. So it was like, Okay, well, like if I want to get an 85 00:03:40.340 --> 00:03:42.900 internship, I need to use a language that other people 86 00:03:42.900 --> 00:03:45.620 use. And I asked people in the 87 00:03:45.620 --> 00:03:48.440 startup community the developer meetups what do people actually 88 00:03:48.440 --> 00:03:51.330 use. And it came back that the community in 89 00:03:51.330 --> 00:03:54.250 Athens had settled on Ruby and JavaScript with 90 00:03:54.250 --> 00:03:57.110 JavaScript being more common more available 91 00:03:57.110 --> 00:04:00.000 easier to get roles with. So I decided. All right, well, I gotta 92 00:04:00.000 --> 00:04:02.890 learn. JavaScript and I picked up a book called Eloquent 93 00:04:02.890 --> 00:04:05.840 JavaScript. And that book has right 94 00:04:05.840 --> 00:04:08.710 next to each other. Two chapters, one on functional programming 95 00:04:08.720 --> 00:04:10.850 and one on object oriented programming. 96 00:04:11.540 --> 00:04:14.470 So I read the functional programming chapter, and it was mind 97 00:04:14.470 --> 00:04:17.160 blowing and confusing, but seemed pretty cool. And I kind of 98 00:04:17.160 --> 00:04:20.070 understood it. And then I read the object oriented chapter 99 00:04:20.080 --> 00:04:23.060 and it was mind blowing and confusing, and I didn't really feel like 100 00:04:23.060 --> 00:04:25.790 I understood it all. I'm sure 101 00:04:25.800 --> 00:04:28.650 most of that it comes down to javascript, not at the 102 00:04:28.650 --> 00:04:31.260 time having a good oriented story. 103 00:04:31.270 --> 00:04:34.160 So The prototypes were very weird for me to 104 00:04:34.160 --> 00:04:36.940 understand. Um, but I 105 00:04:36.940 --> 00:04:39.470 decided. Okay, well, it seems like they're these 106 00:04:39.470 --> 00:04:41.960 paradigms. I gotta learn the most object 107 00:04:41.960 --> 00:04:44.650 oriented programming language. And I gotta learn the most 108 00:04:44.650 --> 00:04:46.560 functional, functional programming language. 109 00:04:47.540 --> 00:04:50.410 And I looked around and I saw Haskell, and 110 00:04:50.410 --> 00:04:53.280 everyone kind of decided that Haskell was the functional 111 00:04:53.280 --> 00:04:56.060 programming language. I was like, All right, well, I'm gonna learn that it's gonna 112 00:04:56.060 --> 00:04:57.650 expand my brain. It'll be cool. 113 00:04:58.240 --> 00:05:01.060 >> Yeah. And then, I guess for the most 114 00:05:01.060 --> 00:05:04.010 object oriented you went ruby slash small talk. 115 00:05:04.020 --> 00:05:06.800 >> Yeah. Ruby, I haven't actually ever worked with small talk. 116 00:05:06.800 --> 00:05:09.710 I've read about it. I read the spec. 117 00:05:09.720 --> 00:05:11.470 I tried implementing it in 118 00:05:12.610 --> 00:05:15.600 rust, but it turns out that's really hard. And so I never actually 119 00:05:15.600 --> 00:05:18.280 got that far with, but yeah, I 120 00:05:18.280 --> 00:05:20.880 decided to do Ruby. Ruby was kind of an easy one to pick, 121 00:05:20.890 --> 00:05:23.690 because, like everyone in the community that was doing 122 00:05:23.690 --> 00:05:25.750 startups wanted to program in Ruby. 123 00:05:26.340 --> 00:05:29.260 >> Yeah, I had a similar experience. Uh, not in 124 00:05:29.260 --> 00:05:31.980 Athens, but in Dallas, Texas. Um, 125 00:05:31.990 --> 00:05:34.990 although I'm surprised there weren't any Java jobs that 126 00:05:34.990 --> 00:05:37.850 seems like kind of the safe bet. But there were definitely a 127 00:05:37.850 --> 00:05:40.730 lot of ruby jobs very popular with the startup community. Ruby 128 00:05:40.730 --> 00:05:43.640 and rails made that type of work. Like you said, 129 00:05:43.640 --> 00:05:46.460 write A Twitter clone. You know, it's just tutorial stuff. 130 00:05:48.740 --> 00:05:51.700 >> Yeah, I think I think the main reason there weren't Java jobs is that 131 00:05:51.700 --> 00:05:54.700 Athens is just like a tiny community, only about 132 00:05:54.700 --> 00:05:57.440 100,000 people in the city. And so there just was 133 00:05:57.440 --> 00:06:00.430 not a tech scene. I think they were, like, maybe 134 00:06:00.430 --> 00:06:03.320 for companies Total. And 135 00:06:03.320 --> 00:06:06.260 there's, like, one or two. There's I think 136 00:06:06.260 --> 00:06:09.230 two tech companies based out of Athens right now that are, like, 137 00:06:09.240 --> 00:06:10.930 modestly successful. 138 00:06:10.940 --> 00:06:13.870 >> Yeah, And they're not the big ones like Oracle or 139 00:06:13.880 --> 00:06:15.790 Cisco. Something like that? 140 00:06:15.800 --> 00:06:18.230 >> No, they might have 100 employees. 141 00:06:18.240 --> 00:06:18.650 >> Yeah, 142 00:06:19.290 --> 00:06:22.200 >> So, Matt, are you telling me you're a bulldog? Is that what you're 143 00:06:22.200 --> 00:06:25.120 telling me here? If you're going to college in 144 00:06:25.120 --> 00:06:25.950 Athens, then 145 00:06:25.960 --> 00:06:28.760 >> I never went to a football 146 00:06:28.760 --> 00:06:31.580 game. People tell me that I'm going to really regret that for the 147 00:06:31.580 --> 00:06:34.460 rest of my life. And I still haven't ever felt any 148 00:06:34.460 --> 00:06:36.040 desire to go to a football game. 149 00:06:36.050 --> 00:06:38.530 >> I'm with you. There Matt. I went to UT I was a 150 00:06:38.530 --> 00:06:41.510 longhorn. I went to one football game just to say that I 151 00:06:41.510 --> 00:06:43.560 had gone and it was okay. 152 00:06:43.640 --> 00:06:46.430 >> I'm a diehard gator. So, 153 00:06:46.440 --> 00:06:49.160 you know, for me Bulldogs, you know, 154 00:06:49.840 --> 00:06:52.340 they're just somebody to beat every year. But, you know, you 155 00:06:52.340 --> 00:06:55.170 ever Matt if you ever find yourself in Florida in 156 00:06:55.170 --> 00:06:57.740 October, you know, maybe we can 157 00:06:57.750 --> 00:07:00.580 go to the Florida Georgia game in Jacksonville is pretty 158 00:07:00.580 --> 00:07:03.460 sweet. That that would be one game that would be 159 00:07:04.040 --> 00:07:07.020 cool to go to just because you're in a Pro Stadium. 160 00:07:07.030 --> 00:07:09.850 They literally split the field half and half like 161 00:07:09.860 --> 00:07:12.440 as fans. So it's kind of neat to see 162 00:07:12.450 --> 00:07:14.030 uh, but it's also 163 00:07:14.030 --> 00:07:16.000 >> I wouldn't wanna be right on that dividing line. 164 00:07:16.010 --> 00:07:18.760 >> Yeah, it's very vicious on that line, 165 00:07:19.140 --> 00:07:22.080 but any who its back on the you know, it's almost 166 00:07:22.080 --> 00:07:24.980 like the divide of, like functional programming and 167 00:07:24.980 --> 00:07:26.590 object oriented programming, you know, 168 00:07:26.600 --> 00:07:28.010 >> We need to fill up a stadium 169 00:07:29.120 --> 00:07:31.380 >> Which one is functional programming? Which one is object oriented 170 00:07:31.390 --> 00:07:32.160 programming? 171 00:07:33.140 --> 00:07:35.390 >> Obviously, the Gators 172 00:07:35.400 --> 00:07:38.210 are the functional programmers just 173 00:07:38.210 --> 00:07:41.140 kidding well, I think we take the teams out of that 174 00:07:41.140 --> 00:07:44.060 one we Let that be more of the the 175 00:07:44.070 --> 00:07:46.990 paradigms that we're really comparing there no, no football teams 176 00:07:46.990 --> 00:07:47.360 needed. 177 00:07:49.390 --> 00:07:49.850 >> Of course 178 00:07:50.240 --> 00:07:53.180 >> Yeah, but awesome. Thanks for, uh, kind of 179 00:07:53.190 --> 00:07:56.070 sharing that experience. So going you 180 00:07:56.070 --> 00:07:58.690 went to the two extremes, which 181 00:07:58.690 --> 00:08:01.580 is, you know, pretty cool. I got I 182 00:08:01.580 --> 00:08:04.310 got started in the middle with, you know, javascript. So it was like, 183 00:08:04.320 --> 00:08:06.360 which one do you pick? Because you can pick either 184 00:08:07.040 --> 00:08:08.030 and good luck 185 00:08:08.040 --> 00:08:10.230 >> or both. Same time. 186 00:08:10.230 --> 00:08:12.450 >> You can just Sprinkle them in 187 00:08:13.700 --> 00:08:16.580 >> Matt. Clearly the functional side kind of won 188 00:08:16.580 --> 00:08:19.090 out. Um, I know it's a really complicated 189 00:08:19.090 --> 00:08:21.890 question, but do you have a feeling of why 190 00:08:21.890 --> 00:08:24.460 you gravitated toward functional rather than object oriented? 191 00:08:25.440 --> 00:08:28.430 >> Well, yeah. So when I was working in that in the 192 00:08:28.430 --> 00:08:31.310 Ruby internship, I ended up finding that it was 193 00:08:31.310 --> 00:08:33.950 easier for me to, like, sketch out a 194 00:08:33.960 --> 00:08:36.750 design for a problem in Haskell. 195 00:08:36.760 --> 00:08:39.370 And I'd come up with ADTs and come with the basic data 196 00:08:39.370 --> 00:08:42.320 transformations. I'd write that out in Haskell, and then I port it all to 197 00:08:42.320 --> 00:08:45.260 Ruby, and the code that I wrote 198 00:08:45.640 --> 00:08:48.050 that was based on, like, functional programming 199 00:08:48.050 --> 00:08:50.810 paradigms in Ruby ended up being like, 200 00:08:50.820 --> 00:08:53.650 easy to test, really easy to work with. It was very 201 00:08:53.650 --> 00:08:56.160 easy to write. You know, TDD approaches 202 00:08:56.740 --> 00:08:58.560 to this highly functional code 203 00:08:59.140 --> 00:09:01.970 and by thinking more along the 204 00:09:01.970 --> 00:09:04.950 lines of data structures and less along the lines of classes and 205 00:09:04.950 --> 00:09:07.820 objects and messages, it was a lot easier to like 206 00:09:07.830 --> 00:09:10.730 model problems such that the solutions 207 00:09:10.730 --> 00:09:13.660 just kind of fell out naturally. One of 208 00:09:13.660 --> 00:09:15.970 the things that we built at that company was 209 00:09:15.970 --> 00:09:18.930 a message tree so that, like 210 00:09:18.930 --> 00:09:21.840 you could have a text message conversation, and based on the answers, you gave 211 00:09:21.850 --> 00:09:24.460 back different actions. would happened on the back end? Uh, 212 00:09:24.470 --> 00:09:27.330 the first draft of this was to have an object with mutable 213 00:09:27.330 --> 00:09:30.240 state that would update its internal state 214 00:09:30.240 --> 00:09:33.230 based on what text messages were received back this 215 00:09:33.230 --> 00:09:35.860 ended up being really complicated and difficult to work with. 216 00:09:36.440 --> 00:09:39.280 So I replaced it with kind of a trie or a 217 00:09:39.280 --> 00:09:41.810 tree of possible messages. And 218 00:09:41.820 --> 00:09:44.750 each level in the tree had a choice 219 00:09:45.440 --> 00:09:48.170 of messages that it could respond to. And 220 00:09:48.170 --> 00:09:50.890 this is extremely easy to express as a 221 00:09:50.890 --> 00:09:53.890 nested, sum type and you 222 00:09:53.890 --> 00:09:56.640 port that over to Ruby and then the whole thing just works really 223 00:09:56.640 --> 00:09:59.250 simply and nicely. Um 224 00:09:59.940 --> 00:10:02.840 and so I just kept finding that I was more and more interested 225 00:10:02.840 --> 00:10:05.710 in functional programming. I never, ever 226 00:10:05.720 --> 00:10:08.660 modeled a problem in Ruby and then tried to solve 227 00:10:08.660 --> 00:10:11.380 it in Haskell. And that's partially because I wasn't solving 228 00:10:11.380 --> 00:10:13.860 problems in Haskell. I didn't have a Haskell job, 229 00:10:14.440 --> 00:10:17.020 but then I got a Haskell 230 00:10:17.020 --> 00:10:19.800 internship. My last year I was working with 231 00:10:19.850 --> 00:10:22.060 Layer three communications out of Atlanta, 232 00:10:22.540 --> 00:10:25.420 and I was getting to use Haskell to solve actual 233 00:10:25.420 --> 00:10:28.090 problems. And I never once 234 00:10:28.100 --> 00:10:30.460 tried to model a problem in Ruby 235 00:10:31.040 --> 00:10:32.590 and then put it into Haskell. 236 00:10:33.140 --> 00:10:34.250 >> Yeah, that's telling. 237 00:10:34.640 --> 00:10:36.250 >> Yeah, it's funny. 238 00:10:37.440 --> 00:10:40.340 >> I just got more and more interested and excited about it. I 239 00:10:40.340 --> 00:10:43.220 started giving talks at the Athens developer meet Up 240 00:10:43.220 --> 00:10:46.170 Group about it and eventually one of the 241 00:10:46.170 --> 00:10:49.050 local startups in Athens actually started building 242 00:10:49.050 --> 00:10:51.260 out some services in Haskell is an experiment, 243 00:10:51.640 --> 00:10:54.110 and it came time for me to 244 00:10:54.110 --> 00:10:56.900 graduate. I talked with the company. I had a few 245 00:10:56.900 --> 00:10:59.620 folks that knew me from the community and they were gonna vouch for 246 00:10:59.620 --> 00:11:02.560 me. And we're like, All right, let's try this out Let's try to 247 00:11:02.560 --> 00:11:05.470 replace more and more of our PHP code with 248 00:11:05.470 --> 00:11:05.860 Haskell. 249 00:11:06.440 --> 00:11:07.180 >> Nice 250 00:11:07.590 --> 00:11:09.070 >> kind of been doing it ever since. 251 00:11:10.040 --> 00:11:12.980 >> Cool. Yeah, That's, uh, that's 252 00:11:12.980 --> 00:11:15.290 really interesting to me because I feel I have some of the 253 00:11:15.290 --> 00:11:17.850 same, um, like I programmed 254 00:11:17.850 --> 00:11:20.600 in scripting languages. I would say PHP, Pearl, 255 00:11:20.600 --> 00:11:23.060 Ruby, that type of stuff for quite a while. 256 00:11:23.070 --> 00:11:26.030 And when I found Haskell, I was really excited because 257 00:11:26.040 --> 00:11:28.960 it felt like the kind of natural 258 00:11:28.960 --> 00:11:31.390 extension of my brain. Like the way that I model 259 00:11:31.390 --> 00:11:34.120 problems matches how 260 00:11:34.120 --> 00:11:36.720 Haskell wants me to model them anyway. So perfect 261 00:11:36.720 --> 00:11:39.610 match and then. You get all these other auxiliary 262 00:11:39.610 --> 00:11:42.610 benefits like you mentioned, they're easier to test. Easier to inspect, 263 00:11:42.620 --> 00:11:45.380 Easier to think about. So yeah, 264 00:11:45.390 --> 00:11:47.560 Curious to have that parallel 265 00:11:48.140 --> 00:11:49.060 track going on? 266 00:11:49.540 --> 00:11:52.350 >> Yeah, I think with I think with Haskell. The thing that's 267 00:11:52.350 --> 00:11:55.300 really incredible is that it has sum types and then 268 00:11:55.300 --> 00:11:56.650 you're able to say or 269 00:11:57.340 --> 00:12:00.250 because in every other programming language, except 270 00:12:00.260 --> 00:12:02.860 until very recently, you could only say and 271 00:12:04.170 --> 00:12:07.110 and I don't know, like, I have never, ever tried to 272 00:12:07.120 --> 00:12:09.360 limit myself to only saying, and 273 00:12:10.240 --> 00:12:13.020 in like my personal life ot the 274 00:12:13.020 --> 00:12:15.510 communications that I use, But it would be so 275 00:12:15.510 --> 00:12:18.480 annoying. I can only imagine how annoying it would 276 00:12:18.490 --> 00:12:21.150 be and then That's kind of how we restrict ourselves in our 277 00:12:21.150 --> 00:12:24.060 programming environments we're saying All right, well, you're only allowed to say, and 278 00:12:24.540 --> 00:12:27.000 you can't say or unless you're literally 279 00:12:27.000 --> 00:12:28.760 saying true or false 280 00:12:29.140 --> 00:12:31.980 >> right only for this one data 281 00:12:31.980 --> 00:12:34.800 type, can you say or Yeah, Cam and I were 282 00:12:34.800 --> 00:12:37.320 talking about this either last week or a couple weeks ago. 283 00:12:37.330 --> 00:12:40.280 Where once you have those tools 284 00:12:40.290 --> 00:12:43.110 of sum types in general being able to say this or 285 00:12:43.110 --> 00:12:45.630 that, you start to wonder like, how did I 286 00:12:45.630 --> 00:12:48.330 accomplish anything before I was able to do this? Just seems 287 00:12:48.330 --> 00:12:50.250 crazy, 288 00:12:51.040 --> 00:12:52.450 >> right? Unless I'm doing some 289 00:12:52.940 --> 00:12:55.580 >> I have kind of the advantage of, like having learned 290 00:12:55.580 --> 00:12:58.530 Haskell within, like having started to learn Haskell within 291 00:12:58.530 --> 00:13:01.460 about, like, six months of learning to program at all. And so 292 00:13:01.460 --> 00:13:04.250 I literally don't know how anyone uses that 293 00:13:04.740 --> 00:13:07.580 I know that there are things like the visitor pattern that you can 294 00:13:07.580 --> 00:13:10.520 do to kind of emulate them. But having never 295 00:13:10.520 --> 00:13:13.520 had to actually write that code, it blows 296 00:13:13.520 --> 00:13:16.290 my mind. It blows my mind that software anywhere 297 00:13:16.290 --> 00:13:19.250 works for a huge amount of reasons, 298 00:13:19.250 --> 00:13:20.800 and that's only one of them. 299 00:13:20.810 --> 00:13:21.650 >> Yes, agreed. 300 00:13:22.040 --> 00:13:25.030 >> and I think for me like my experience. I 301 00:13:25.030 --> 00:13:27.730 started coding in college like that was my first ever 302 00:13:27.730 --> 00:13:30.590 experience, even thinking about programming. And at 303 00:13:30.590 --> 00:13:33.550 that point, you know, you're learning Java or 304 00:13:33.550 --> 00:13:36.230 C plus plus and like it's very 305 00:13:36.230 --> 00:13:39.010 object oriented like there's not a ton of functional 306 00:13:39.010 --> 00:13:41.800 programming concepts at least taught at the program I was 307 00:13:41.800 --> 00:13:44.150 in, Uh, and, like 308 00:13:44.640 --> 00:13:47.250 you know, once I had learned I got a JavaScript 309 00:13:47.250 --> 00:13:49.920 internship, actually hear it ITProTV like that's where I 310 00:13:49.920 --> 00:13:52.510 started. Um, you know, it offered 311 00:13:52.510 --> 00:13:55.050 me, like, a little more flexibility than all the, you know, 312 00:13:55.440 --> 00:13:58.180 object oriented programming languages. I was used to like C plus plus 313 00:13:58.180 --> 00:14:00.930 and Java. Um, but then, 314 00:14:00.940 --> 00:14:03.890 thankfully, a couple years later, I was introduced a Haskell 315 00:14:03.900 --> 00:14:06.380 Well, Elm and then Haskell and like, 316 00:14:06.390 --> 00:14:09.190 yeah, like Taylor said, it's just like you said. 317 00:14:09.190 --> 00:14:12.190 It's awesome to be able to really model things 318 00:14:12.190 --> 00:14:13.860 in such an inclusive way. Like 319 00:14:14.340 --> 00:14:17.330 there's so many ways and things things you can 320 00:14:17.330 --> 00:14:20.140 express with Haskell that you just can't clearly 321 00:14:20.140 --> 00:14:22.930 express in JavaScript or any of those other object 322 00:14:22.930 --> 00:14:24.070 oriented programming languages. 323 00:14:24.190 --> 00:14:26.260 >> So Elm was the gateway drug for you. 324 00:14:26.640 --> 00:14:29.360 >> Oh, yeah, Matt 325 00:14:29.560 --> 00:14:30.660 Have you ever used Elm. 326 00:14:31.510 --> 00:14:32.250 >> I have not. 327 00:14:33.840 --> 00:14:36.360 >> Okay, well, if you're interested in front end languages, 328 00:14:36.840 --> 00:14:37.420 you should try it. 329 00:14:37.420 --> 00:14:40.310 >> It's a good one. It's funny. I feel like the Haskell 330 00:14:40.310 --> 00:14:43.160 community has kind of a complicated relationship with 331 00:14:43.160 --> 00:14:45.810 Elm because on the one hand, for many people it 332 00:14:45.810 --> 00:14:48.550 is a gateway like it's a simple version of 333 00:14:48.550 --> 00:14:50.820 Haskell that can introduce you to the concepts. 334 00:14:50.830 --> 00:14:53.720 But the other side of that double edged 335 00:14:53.730 --> 00:14:56.480 sword is that it doesn't have all the fancy 336 00:14:56.480 --> 00:14:59.480 stuff that Haskell has that people that use Haskell really like 337 00:14:59.490 --> 00:15:01.950 like effect systems and monads and all this stuff. 338 00:15:02.040 --> 00:15:04.870 >> Mhm, yeah, the type class system. Not 339 00:15:04.870 --> 00:15:07.250 having that. does really trip me up 340 00:15:07.840 --> 00:15:10.620 a lot. And then they have one type class. It's called 341 00:15:10.620 --> 00:15:13.270 comparable. And you're like, What? What is 342 00:15:13.270 --> 00:15:13.930 this 343 00:15:14.480 --> 00:15:15.160 >> Magic! 344 00:15:15.640 --> 00:15:18.410 >> Yeah, I'm familiar with 345 00:15:18.410 --> 00:15:21.410 Elm. It seems like a really like it has the most important 346 00:15:21.410 --> 00:15:24.370 features for doing functional programming correctly. Like I wouldn't 347 00:15:24.370 --> 00:15:27.360 want it. I wouldn't want to take away sum types I wouldn't want 348 00:15:27.360 --> 00:15:30.300 to take away, like pure functions on. I wouldn't want to 349 00:15:30.300 --> 00:15:32.820 take away like generics or type variables or whatever they 350 00:15:32.820 --> 00:15:35.540 have, because I feel like they have like the most 351 00:15:35.540 --> 00:15:38.260 important things for it. Um, 352 00:15:39.240 --> 00:15:39.710 yeah, 353 00:15:39.720 --> 00:15:41.660 >> it's kind of like the simplest possible. 354 00:15:42.040 --> 00:15:45.010 >> Yes, yes, My take on the front end 355 00:15:45.010 --> 00:15:47.870 is that I don't want to do it. And so 356 00:15:47.870 --> 00:15:50.790 whatever technology will allow me to push that on to someone 357 00:15:50.790 --> 00:15:53.780 else. That's the best technology for me on the 358 00:15:53.780 --> 00:15:56.310 front end. If I had to write my own like 359 00:15:56.310 --> 00:15:59.290 interactive, single page app, I would want to use probably pure 360 00:15:59.290 --> 00:16:02.290 script. My experiences with that have always 361 00:16:02.290 --> 00:16:04.940 just been extremely positive. Great 362 00:16:04.940 --> 00:16:07.830 community, great language. Libraries are pretty 363 00:16:07.830 --> 00:16:09.890 high quality when they exist. And 364 00:16:09.890 --> 00:16:10.920 >> the editor tooling 365 00:16:10.920 --> 00:16:13.360 >> Yes, the editor tooling. It just works. It's amazing. 366 00:16:13.740 --> 00:16:16.670 >> Yeah, Haskell is getting there, but it's a 367 00:16:16.670 --> 00:16:17.960 slow process. 368 00:16:17.970 --> 00:16:20.710 >> Yeah, you know, I dedicate about eight 369 00:16:20.710 --> 00:16:23.440 hours a year to trying to figure out if the editor 370 00:16:23.440 --> 00:16:26.390 tooling situation has gotten any better, and it 371 00:16:26.400 --> 00:16:29.340 has not every single year that 372 00:16:29.340 --> 00:16:30.160 I've tried to do it. 373 00:16:30.740 --> 00:16:33.510 >> Well, this year might be the one that it changes the year of the Linux 374 00:16:33.510 --> 00:16:35.940 desktop and the Haskell tooling 375 00:16:35.950 --> 00:16:37.030 coincide. 376 00:16:37.040 --> 00:16:38.660 >> I would love it. 377 00:16:39.040 --> 00:16:41.200 >> Mhm. That's awesome. 378 00:16:41.210 --> 00:16:44.140 >> my problem Is that it always. It always works fine on small 379 00:16:44.140 --> 00:16:46.740 projects. It does seem like the 380 00:16:46.740 --> 00:16:49.580 size of a project is increasing 381 00:16:49.580 --> 00:16:51.720 before the tooling kind of falls over. 382 00:16:51.730 --> 00:16:54.060 But I have received 383 00:16:54.440 --> 00:16:57.320 I've tried to get it working with the work project at Lumi 384 00:16:57.330 --> 00:17:00.020 and it does not work. What 385 00:17:00.020 --> 00:17:02.920 happens is it will get through, like, kind of the beginning of the 386 00:17:02.920 --> 00:17:05.860 module hierarchy. alright so the stuff that 387 00:17:05.860 --> 00:17:08.760 doesn't have any dependencies, you'll get pretty good feedback. But once we 388 00:17:08.760 --> 00:17:11.680 start getting into like, oh, I don't know, 389 00:17:11.680 --> 00:17:14.320 like the we have about 700 modules in the 390 00:17:14.320 --> 00:17:17.290 project. And once you start getting into about module 300 391 00:17:17.290 --> 00:17:19.810 400 the feedback just takes 392 00:17:19.820 --> 00:17:22.380 way too long for it to be useful. 393 00:17:22.390 --> 00:17:24.630 >> So what we need is the, 394 00:17:24.640 --> 00:17:27.620 uh size that the tooling is okay with 395 00:17:27.630 --> 00:17:30.220 to grow faster than the average size of a Haskell 396 00:17:30.220 --> 00:17:30.780 project. 397 00:17:30.790 --> 00:17:32.160 >> That's correct. 398 00:17:32.170 --> 00:17:32.950 >> It'll tip over eventually. 399 00:17:33.840 --> 00:17:36.640 >> I can write Haskell code faster than people can release. 400 00:17:36.640 --> 00:17:38.730 Fast tooling. Unfortunately. 401 00:17:38.740 --> 00:17:41.550 >> Just slow down Matt. The problem will take care of itself. 402 00:17:41.940 --> 00:17:44.890 Write less code. Um, well, cool. 403 00:17:44.890 --> 00:17:47.520 I appreciate you, You know, giving us some insight in 404 00:17:47.520 --> 00:17:50.010 to How you got to where you are. Um, I 405 00:17:50.010 --> 00:17:52.890 wanna turn our attention a little bit now. toward the book 406 00:17:52.890 --> 00:17:55.660 that you're in the process of writing. I know that you've been 407 00:17:55.670 --> 00:17:58.670 working on it for at least several months now, and you have 408 00:17:58.670 --> 00:18:01.590 several hundred pages to show for yourself. But can 409 00:18:01.590 --> 00:18:03.170 you, uh, tell us about it? 410 00:18:03.640 --> 00:18:06.510 >> Yeah, absolutely. Um, the book actually 411 00:18:06.510 --> 00:18:09.440 started out as a joint project between me and Sandy 412 00:18:09.440 --> 00:18:12.100 Maguire. He was going to 413 00:18:12.110 --> 00:18:14.830 write all the advanced type level 414 00:18:14.830 --> 00:18:17.780 system stuff, and I was going to write another section 415 00:18:17.780 --> 00:18:20.260 on other advanced Haskell techniques. 416 00:18:20.840 --> 00:18:23.050 And I was, 417 00:18:23.200 --> 00:18:26.170 uh, unable to work on 418 00:18:26.170 --> 00:18:29.040 the book fast enough. And so he ended up taking his half of it and 419 00:18:29.040 --> 00:18:32.010 releasing thinking with types. So I've actually I've 420 00:18:32.010 --> 00:18:34.960 had this project on the back burner for a really long time, but 421 00:18:34.960 --> 00:18:37.550 I've had a hard time getting started with it until 422 00:18:37.940 --> 00:18:40.900 really Until about August of last year, I was diagnosed with 423 00:18:40.900 --> 00:18:43.850 ADHD And started receiving treatment for it. And in 424 00:18:43.850 --> 00:18:46.370 the intervening time, I'm up to about 400 425 00:18:47.020 --> 00:18:49.670 I think 418 pages. 426 00:18:50.040 --> 00:18:50.460 >> Wow. 427 00:18:50.840 --> 00:18:53.800 >> Yeah, yeah, it's amazing what executive dysfunction 428 00:18:53.800 --> 00:18:55.510 can do or not, 429 00:18:55.520 --> 00:18:56.310 >> So the treatment is working 430 00:18:57.240 --> 00:19:00.190 >> Yeah, but the idea initially 431 00:19:00.190 --> 00:19:02.610 I wanted to write kind of a book on intermediate to Advanced 432 00:19:02.610 --> 00:19:05.500 Haskell that would be useful for someone that had completed 433 00:19:05.500 --> 00:19:08.050 something like Haskell Book or Haskell from First 434 00:19:08.050 --> 00:19:11.010 Principles or Graham Hutton's book or some other 435 00:19:11.010 --> 00:19:13.900 introductory text, and was stuck knowing 436 00:19:13.900 --> 00:19:16.700 the language but not knowing how to actually build 437 00:19:16.700 --> 00:19:19.680 something useful with Haskell. Because once you 438 00:19:19.680 --> 00:19:22.050 understand monads and maybe monad Transformers, 439 00:19:22.440 --> 00:19:25.120 you're able to, like, write, code and 440 00:19:25.120 --> 00:19:27.600 understand code. But then you're kind of plopped into this 441 00:19:27.600 --> 00:19:29.760 ecosystem. That's mystifying. 442 00:19:30.140 --> 00:19:32.750 That is just very difficult to 443 00:19:32.750 --> 00:19:35.550 discover, despite the presence of tools like Hoogle, 444 00:19:35.940 --> 00:19:38.550 which, given a type signature, could give or, 445 00:19:38.940 --> 00:19:41.770 you know, some things that will implement that type 446 00:19:41.770 --> 00:19:44.610 signature. And so 447 00:19:44.610 --> 00:19:47.220 this book. I'm really I'm wanting to say 448 00:19:47.230 --> 00:19:50.230 you can use Haskell in industry. You could make a 449 00:19:50.230 --> 00:19:52.660 lot of money with it, and it's really cool. 450 00:19:52.670 --> 00:19:55.610 But it's tricky. It's different than other systems. 451 00:19:55.620 --> 00:19:57.860 It's different than other ecosystems. 452 00:19:58.240 --> 00:20:01.150 So what I want to do here is say, I don't 453 00:20:01.150 --> 00:20:03.620 know like what the best way is, but I have 454 00:20:03.620 --> 00:20:06.460 watched projects work and I've watched projects 455 00:20:06.460 --> 00:20:09.220 fail, and this is kind of the 456 00:20:09.220 --> 00:20:12.050 advice that I have to have your project work. 457 00:20:13.440 --> 00:20:15.830 >> I mean, I think, honestly, the community needs that 458 00:20:15.840 --> 00:20:18.820 Yeah, think from as 459 00:20:18.820 --> 00:20:21.780 a programmer myself who you know works in Haskell 460 00:20:21.780 --> 00:20:24.210 day to day. And it's maintaining, you know, a 461 00:20:24.220 --> 00:20:27.140 relatively large project that continues to grow. 462 00:20:27.150 --> 00:20:29.460 You know, it helps to be able to 463 00:20:30.740 --> 00:20:33.400 kind of say Okay, what can we improve and 464 00:20:33.410 --> 00:20:36.330 why are we going to make these improvements? So I think that would be 465 00:20:36.340 --> 00:20:38.360 really beneficial. 466 00:20:39.240 --> 00:20:41.850 >> Yeah, And the Haskell community has a lot of 467 00:20:41.860 --> 00:20:44.590 beginner. Resources like you mentioned, you 468 00:20:44.590 --> 00:20:46.770 know, Haskell programming from First Principles, 469 00:20:46.780 --> 00:20:49.580 Hutton's book and the other one, 470 00:20:49.590 --> 00:20:51.600 um, learn you a Haskell for great good 471 00:20:51.610 --> 00:20:54.020 and there's Also, a lot of like 472 00:20:54.030 --> 00:20:56.930 very expert things 473 00:20:56.930 --> 00:20:59.550 like real world Haskell or like you mentioned Sandy's 474 00:20:59.550 --> 00:21:02.550 book and many, many others, Um and So I 475 00:21:02.550 --> 00:21:05.290 feel like you're feeling an important gap here. where like, Yeah, I can do 476 00:21:05.290 --> 00:21:08.170 Haskell in the small But how do I 477 00:21:08.630 --> 00:21:10.670 leverage those skills to produce a system? 478 00:21:11.640 --> 00:21:14.550 >> Yeah, absolutely. And I think 479 00:21:14.550 --> 00:21:17.080 one of the one of the difficult things about 480 00:21:17.090 --> 00:21:19.480 Haskell is that its attributes as a 481 00:21:19.480 --> 00:21:22.250 language. It's so much more productive 482 00:21:22.260 --> 00:21:24.660 when it comes to just the act of writing software. 483 00:21:25.140 --> 00:21:27.840 That part of being a Haskell developer is 484 00:21:27.840 --> 00:21:30.170 so much better. And 485 00:21:30.740 --> 00:21:33.710 I am obviously biased, uh, in part 486 00:21:33.710 --> 00:21:36.530 because I use Haskell all the time and also in part because I 487 00:21:36.530 --> 00:21:39.130 don't actually know what it's really 488 00:21:39.130 --> 00:21:42.000 like to be a professional engineer with any other 489 00:21:42.000 --> 00:21:44.970 kind of language. But 490 00:21:44.980 --> 00:21:47.950 I do think that Haskell is a superpower in terms of software 491 00:21:47.950 --> 00:21:50.830 engineering, but that's only one 492 00:21:50.830 --> 00:21:53.730 part of the system. And so when one 493 00:21:53.730 --> 00:21:56.630 part of a system becomes super optimized and super 494 00:21:56.630 --> 00:21:59.360 fast, you have to let the rest of the 495 00:21:59.360 --> 00:22:01.550 system kind of optimize around this new set point. 496 00:22:02.740 --> 00:22:05.620 It's kind of like, you know, when you're writing concurrency 497 00:22:05.620 --> 00:22:08.300 code, if you have a queue and that queue doesn't have 498 00:22:08.300 --> 00:22:11.000 a limit in its size or it doesn't have any back 499 00:22:11.000 --> 00:22:13.850 pressure, that queue can overflow and blow everything up. 500 00:22:14.440 --> 00:22:14.920 >> Yeah, 501 00:22:14.930 --> 00:22:17.920 >> um, at the at the first job, we actually witnessed 502 00:22:17.920 --> 00:22:19.360 this kind of 503 00:22:20.440 --> 00:22:23.290 in the actual performance of the Haskell 504 00:22:23.290 --> 00:22:25.960 code and also in how the team was working. 505 00:22:26.540 --> 00:22:29.400 I think I was able to mostly on my 506 00:22:29.400 --> 00:22:32.370 own, develop a code base that was around 35 to 507 00:22:32.370 --> 00:22:33.770 40,000 lines of Haskell. 508 00:22:34.240 --> 00:22:36.160 >> Wow, What kind of time frame? 509 00:22:36.740 --> 00:22:39.560 >> Um, I think that was within my first like 510 00:22:39.570 --> 00:22:41.760 year of working there. 511 00:22:42.740 --> 00:22:45.450 Yeah, it was pretty early, but you know, 512 00:22:45.450 --> 00:22:48.300 Haskell, it's just so easy to write code that works, that it 513 00:22:48.300 --> 00:22:51.180 was just for the most part. Also, it 514 00:22:51.180 --> 00:22:54.040 helps that I was to some extent taking existing 515 00:22:54.040 --> 00:22:56.260 PHP projects and transpiling or 516 00:22:56.640 --> 00:22:59.540 porting, transcribing them porting them, making them better. 517 00:22:59.550 --> 00:23:02.500 But we had one system that I 518 00:23:02.500 --> 00:23:05.170 wrote, and it was so much 519 00:23:05.170 --> 00:23:08.170 faster than the equivalent PHP system that 520 00:23:08.170 --> 00:23:10.890 when we dropped the new system in it revealed a 521 00:23:10.890 --> 00:23:13.760 lot of performance problems in every other 522 00:23:13.760 --> 00:23:16.360 component that touched it. We had servers that were 523 00:23:16.360 --> 00:23:18.750 falling over because they weren't used to having 524 00:23:18.760 --> 00:23:21.640 a, you know, back in server that was responding so 525 00:23:21.640 --> 00:23:24.470 fast. So we ended up having to 526 00:23:24.470 --> 00:23:27.330 actually artificially slow down the Haskell service in 527 00:23:27.330 --> 00:23:30.130 order to deal with that problem. Now, eventually, a 528 00:23:30.130 --> 00:23:33.010 lot more of those services got folded over into Haskell and that 529 00:23:33.010 --> 00:23:36.010 made everything like a lot faster. But you 530 00:23:36.010 --> 00:23:38.920 can't just make one part of a system better and 531 00:23:38.920 --> 00:23:41.050 expect the whole system to improve. 532 00:23:41.840 --> 00:23:44.050 And I think that's something that's, like not, 533 00:23:44.540 --> 00:23:47.540 uh, intuitive. It's not something that I think 534 00:23:47.550 --> 00:23:50.430 people are. It's It is weird, 535 00:23:50.440 --> 00:23:53.360 right? Like you expect. Okay, well, I'm gonna make one part faster. Everything else should 536 00:23:53.360 --> 00:23:53.950 get better. 537 00:23:54.440 --> 00:23:57.300 >> Well, or at the very least. Everything else should 538 00:23:57.300 --> 00:24:00.300 not get worse. But like you, you know, like you lay 539 00:24:00.300 --> 00:24:02.800 out in this example, things kind of depend 540 00:24:02.800 --> 00:24:05.750 on a particular level of performance. So if you just completely 541 00:24:05.750 --> 00:24:08.570 blow out of the water, they don't know what to do anymore, or they 542 00:24:08.570 --> 00:24:10.010 fall over, or who knows? 543 00:24:10.020 --> 00:24:12.970 >> Yeah. And on the productivity side of 544 00:24:12.970 --> 00:24:15.730 things, you know, you might 545 00:24:15.740 --> 00:24:18.360 you might be a 10x developer if you're using 546 00:24:18.360 --> 00:24:21.110 Haskell. But that only applies to the actual code that you 547 00:24:21.110 --> 00:24:24.070 write. You have to write documentation 548 00:24:24.080 --> 00:24:26.980 you have to care about, like, observability of a feature you have to care 549 00:24:26.980 --> 00:24:29.960 about, You know, a continuous integration and testing and 550 00:24:29.960 --> 00:24:32.690 all this stuff. And those aren't any faster with 551 00:24:32.690 --> 00:24:35.170 Haskell. In fact, I think in a lot of cases they're slower, 552 00:24:36.040 --> 00:24:38.990 Um, integrating with like, an error reporting service 553 00:24:39.000 --> 00:24:41.780 is something that I've had to do recently and in every 554 00:24:41.780 --> 00:24:44.470 other language you just like import error 555 00:24:44.470 --> 00:24:46.770 reporting service into your package 556 00:24:46.770 --> 00:24:49.550 file. Crazy java metadata 557 00:24:49.550 --> 00:24:52.110 Hacking happens after that. Or like, Ruby 558 00:24:52.120 --> 00:24:55.080 scripts get rewritten behind the scenes and you just get error 559 00:24:55.080 --> 00:24:56.180 reporting. Haskell. 560 00:24:56.180 --> 00:24:58.200 >> None of it is pretty, but it just works. 561 00:24:58.210 --> 00:25:01.150 >> Yeah, exactly. And, you know, in Haskell that's 562 00:25:01.150 --> 00:25:03.780 not the case. Um, nothing 563 00:25:03.790 --> 00:25:06.520 is off the shelf. Nothing just works. You have to 564 00:25:06.520 --> 00:25:09.360 thread in everything that you want to care about. 565 00:25:09.740 --> 00:25:12.620 Um I think that that does over the 566 00:25:12.620 --> 00:25:15.460 long run give you, like, a better signal to noise ratio. 567 00:25:16.040 --> 00:25:18.630 But, you know, it's still 568 00:25:18.630 --> 00:25:21.420 annoying that we have to put together our own stack 569 00:25:21.420 --> 00:25:23.470 traces for exceptions. 570 00:25:23.480 --> 00:25:26.390 >> Oh, my gosh, you're striking a nerve 571 00:25:26.390 --> 00:25:29.150 here because we've been struggling with this at our job as well 572 00:25:29.160 --> 00:25:32.160 as, well, where we want to just plug in exception reporting. 573 00:25:32.240 --> 00:25:35.160 But we have to remember like, Oh, we forked a thread here, 574 00:25:35.160 --> 00:25:38.060 but we forgot to install the exception handler. So those have just been 575 00:25:38.240 --> 00:25:39.970 throwing into the ether in the meantime. 576 00:25:40.440 --> 00:25:41.160 >> Mhm. 577 00:25:41.740 --> 00:25:44.680 >> Yeah, I I've It's a problem that I keep coming back 578 00:25:44.680 --> 00:25:47.570 to. I want to have, a exception wrapper 579 00:25:47.570 --> 00:25:50.090 type that carries just a list 580 00:25:50.100 --> 00:25:52.990 of like I don't know, Aeson values. Just 581 00:25:53.000 --> 00:25:55.660 some random metadata. You can attach an exception, 582 00:25:55.940 --> 00:25:58.250 but it is so 583 00:25:58.250 --> 00:26:01.050 tricky to figure out how to 584 00:26:01.060 --> 00:26:04.060 deal with that information. Because if you 585 00:26:04.060 --> 00:26:06.990 want to do this and you want to attach that information directly to the 586 00:26:06.990 --> 00:26:09.940 exceptions, well, now you have 587 00:26:09.940 --> 00:26:12.470 to write a special catch function that 588 00:26:12.470 --> 00:26:15.390 looks at the exception and tries to determine. Okay, Well, 589 00:26:15.390 --> 00:26:17.900 are you what I'm trying to catch? Or are you a 590 00:26:17.910 --> 00:26:20.880 version that's annotated with metadata? And I need to 591 00:26:20.880 --> 00:26:23.050 preserve that metadata if the exception gets 592 00:26:23.050 --> 00:26:24.060 rethrown? 593 00:26:25.240 --> 00:26:28.100 >> We've been struggling. You're hitting all the nerves here because we have this 594 00:26:28.100 --> 00:26:30.830 exact problem, not with arbitrary metadata, but just with the call 595 00:26:30.830 --> 00:26:33.370 stack. And it's, you know, we have to provide the wrapping and 596 00:26:33.370 --> 00:26:36.350 unwrapping and catching and throwing and all this. It's 597 00:26:36.350 --> 00:26:39.310 painful and, you know, bringing it back to the larger point. 598 00:26:39.320 --> 00:26:41.790 Earlier, we were talking about tooling being bad, and I 599 00:26:41.790 --> 00:26:44.650 think one of the reasons that Haskell tooling 600 00:26:44.650 --> 00:26:47.560 can kind of get away with being subpar 601 00:26:47.640 --> 00:26:50.440 is that it is such a productive language 602 00:26:50.440 --> 00:26:53.290 that you can kind of get away without it until you 603 00:26:53.290 --> 00:26:56.190 start getting little glimpses of it. You know, like we've been trying 604 00:26:56.190 --> 00:26:59.160 out Haskell Language server. And for the few minutes that it 605 00:26:59.160 --> 00:27:02.000 manages to work for, some of us we're like, Oh, 606 00:27:02.000 --> 00:27:04.820 my gosh, this is game changing. I want to do this every 607 00:27:04.820 --> 00:27:07.430 day, but then it crashes it. Sorry, I don't want to 608 00:27:07.430 --> 00:27:10.350 besmirch HLS. It's a fantastic project. We haven't had good 609 00:27:10.350 --> 00:27:11.860 luck getting it to work at work. 610 00:27:12.240 --> 00:27:14.950 >> Yeah, like you said, smaller 611 00:27:14.950 --> 00:27:17.860 projects, it looks works pretty well. It's fantastic. 612 00:27:17.860 --> 00:27:20.350 And then we try to bring it into our larger 613 00:27:21.440 --> 00:27:24.260 project, and we just have trouble. But, you know, we have 614 00:27:24.260 --> 00:27:27.260 hope we're going to continue to try it and 615 00:27:27.420 --> 00:27:28.670 do anything we can to help it. 616 00:27:29.140 --> 00:27:29.670 >> Mhm. 617 00:27:30.140 --> 00:27:32.960 >> Yeah. I do think that you're right and 618 00:27:32.960 --> 00:27:35.700 that the additional productivity of the 619 00:27:35.700 --> 00:27:38.690 language means that we can get away with less. IDE support I 620 00:27:38.690 --> 00:27:41.600 actually have ah kind of tongue in cheek law of the 621 00:27:41.600 --> 00:27:44.210 universe. It's the law of 622 00:27:44.210 --> 00:27:46.300 conservation, of engineering, suffering 623 00:27:46.740 --> 00:27:49.600 where if you make one thing nicer, something else will get 624 00:27:49.600 --> 00:27:52.420 worse in order to make sure that you're always as unhappy as you can 625 00:27:52.420 --> 00:27:52.770 be. 626 00:27:53.240 --> 00:27:55.990 >> Well, yeah, you always gotta aspire to better, 627 00:27:55.990 --> 00:27:58.950 right? You know, it's gonna always have something to work for. 628 00:27:58.960 --> 00:28:01.650 >> This sounds like a glass half full glass, half empty type of 629 00:28:01.650 --> 00:28:03.950 way. 630 00:28:04.440 --> 00:28:06.270 >> Always a glass half full kind of person. 631 00:28:10.240 --> 00:28:13.060 >> Well, awesome. Yeah, 632 00:28:13.070 --> 00:28:15.910 you said so much that I agree with. And I feel like we could spend 633 00:28:15.920 --> 00:28:18.580 30 minutes just talking about each and every little bit of 634 00:28:18.580 --> 00:28:21.570 minutia here. But bringing 635 00:28:21.570 --> 00:28:24.520 it back to your book it sounds like a fascinating resource 636 00:28:24.520 --> 00:28:26.680 and super valuable for the community. 637 00:28:26.690 --> 00:28:29.570 Um, I managed to snag a pre release 638 00:28:29.570 --> 00:28:32.520 version by advertising in Haskell Weekly. 639 00:28:32.530 --> 00:28:35.250 Not everyone can do that. If other people want to get their hands on 640 00:28:35.250 --> 00:28:36.350 this, how can they do it? 641 00:28:37.140 --> 00:28:39.760 >> Well, the book is currently available for sale on Leanpub. 642 00:28:40.240 --> 00:28:41.640 And so, if you, I think to 643 00:28:41.640 --> 00:28:44.760 leanpub.com/production-haskell 644 00:28:44.770 --> 00:28:47.620 that will bring up the book sale page and you'll be able to 645 00:28:47.620 --> 00:28:50.450 acquire it. If you get it now, you will receive 646 00:28:50.450 --> 00:28:53.200 updates. You'll receive email updates. I'm doing a 647 00:28:53.210 --> 00:28:56.150 release about once a month, and I'm including a change log 648 00:28:56.150 --> 00:28:58.800 and everything. I do You can follow the Twitter 649 00:28:58.800 --> 00:29:00.760 account @ProdHaskell 650 00:29:01.540 --> 00:29:04.350 and that will also keep you up to date on 651 00:29:04.350 --> 00:29:06.960 the status of the book and how things were going with it. 652 00:29:07.840 --> 00:29:10.700 >> Excellent. I'll be sure to 653 00:29:10.700 --> 00:29:13.600 include those links in the show notes. So if people 654 00:29:13.600 --> 00:29:16.540 are lazy and looking for something to click on, I'll give 655 00:29:16.540 --> 00:29:18.850 them just that. Yeah. 656 00:29:19.840 --> 00:29:22.670 Um and yeah. Is there anything else you want people to 657 00:29:22.670 --> 00:29:25.670 know about? You like your twitter handle or where 658 00:29:25.670 --> 00:29:26.950 to find you what you're doing? 659 00:29:27.340 --> 00:29:29.090 >> Oh, yeah. My twitter handle is 660 00:29:29.090 --> 00:29:30.760 @MattOfLambda. 661 00:29:31.340 --> 00:29:33.880 Um and 662 00:29:33.880 --> 00:29:36.830 yeah, I'm on the FP chat slack a lot. So if 663 00:29:36.830 --> 00:29:39.660 you have any questions or want help with programming Haskell, 664 00:29:39.660 --> 00:29:42.310 hop on there, we'll be able to get you 665 00:29:42.310 --> 00:29:44.460 sorted, make sure everything is good and happy. 666 00:29:45.740 --> 00:29:46.670 >> Excellent. 667 00:29:47.940 --> 00:29:50.580 >> Yeah, mostly, though, I'm doing a lot of bike riding. 668 00:29:51.240 --> 00:29:54.200 >> Oh, yeah, Let's talk about that because I'm also 669 00:29:54.210 --> 00:29:55.850 a bike rider. Um, 670 00:29:56.440 --> 00:29:59.040 >> This is nowp a cycling podcast. 671 00:29:59.050 --> 00:30:01.630 >> Yeah, maybe we need to spin off 672 00:30:01.640 --> 00:30:04.600 podcast, um, It's funny I was listening to another 673 00:30:04.600 --> 00:30:07.570 podcast today called the bike shed. But it's not about bikes, 674 00:30:07.580 --> 00:30:10.260 it's about programming. 675 00:30:11.040 --> 00:30:14.000 Um, but yeah, it's funny. I think that cycling is one of the more 676 00:30:14.000 --> 00:30:16.900 popular, like the Venn diagram of 677 00:30:16.900 --> 00:30:19.730 cyclists and programmers I think is almost a circle. 678 00:30:19.740 --> 00:30:22.580 >> Yeah, it's interesting. I 679 00:30:22.580 --> 00:30:25.370 like it because it's so It's so 680 00:30:25.370 --> 00:30:28.130 deterministic. I just got an 681 00:30:28.130 --> 00:30:31.020 indoor trainer. Um, that has, like a power meter 682 00:30:31.020 --> 00:30:33.870 built in and training with a 683 00:30:33.870 --> 00:30:36.420 power meter is about as objective is like 684 00:30:36.430 --> 00:30:38.950 lifting weights at the gym and that you know 685 00:30:38.950 --> 00:30:40.300 exactly how much work you're 686 00:30:40.300 --> 00:30:42.670 doing, 687 00:30:42.940 --> 00:30:45.580 man. Like you add the time dimension to 688 00:30:45.580 --> 00:30:48.460 power. And it's like the analysis you can do on it 689 00:30:48.460 --> 00:30:51.360 is nuts. It's fun. 690 00:30:51.360 --> 00:30:54.350 Like I've been trying out all these different apps and checking out 691 00:30:54.350 --> 00:30:56.870 all these different ways of analyzing your fitness and your 692 00:30:56.870 --> 00:30:59.620 freshness and how much you should train. There's 693 00:30:59.620 --> 00:31:02.450 so much there. It's so cool. I spend more 694 00:31:02.450 --> 00:31:04.980 time looking at charts than I do actually riding my bike. Maybe 695 00:31:04.990 --> 00:31:07.750 >> you gotta look at the 696 00:31:07.750 --> 00:31:10.150 charts while you're on the bike, you know, multi 697 00:31:10.150 --> 00:31:12.450 task. 698 00:31:13.340 --> 00:31:13.800 >> That's a valid option 699 00:31:13.810 --> 00:31:16.560 >> that sounds awesome. Mhm. 700 00:31:17.740 --> 00:31:20.620 All right, well, I think that'll about do it for 701 00:31:20.620 --> 00:31:23.520 us. Like I said, there's so much more that we could dig into here. 702 00:31:23.520 --> 00:31:26.380 But I get the feeling reading the book will be a good 703 00:31:26.380 --> 00:31:28.550 opportunity to dig into some of those topics. 704 00:31:28.940 --> 00:31:31.860 Um, so thank you for joining us today, Matt. I hope you had a good 705 00:31:31.860 --> 00:31:33.050 time. I've had a great time. 706 00:31:33.540 --> 00:31:33.810 >> Yeah. 707 00:31:33.820 --> 00:31:34.450 >> Awesome having you on the show 708 00:31:34.450 --> 00:31:35.960 >> this was fun. Thank you so much. 709 00:31:37.130 --> 00:31:39.950 >> And, Cam, you wanna walk us out with the usual 710 00:31:41.420 --> 00:31:44.350 >> where they can find us? Uh, here at Haskell weekly. 711 00:31:44.360 --> 00:31:47.050 You can find us on the 712 00:31:47.060 --> 00:31:48.370 Web at 713 00:31:48.370 --> 00:31:51.070 HaskellWeekly.News. You can also 714 00:31:51.070 --> 00:31:53.520 find us on pretty much any social 715 00:31:53.530 --> 00:31:54.920 media platform 716 00:31:54.930 --> 00:31:57.360 @HaskellWeekly. Um, 717 00:31:57.640 --> 00:32:00.620 and obviously sign up for the newsletter if you already have it. If you haven't 718 00:32:00.620 --> 00:32:03.490 already that way you can see, um, all the 719 00:32:03.490 --> 00:32:06.300 great content that the community is creating and you 720 00:32:06.300 --> 00:32:08.790 know we've We've done covered quite a few 721 00:32:08.800 --> 00:32:11.770 topics that Matt's even talked about. So, you know, if you're ever 722 00:32:11.770 --> 00:32:14.560 interested in what's going on the community, please check out 723 00:32:14.570 --> 00:32:17.490 our and subscribe to the weekly news 724 00:32:17.490 --> 00:32:18.080 letter. 725 00:32:18.090 --> 00:32:20.960 >> That's the important part. And we are brought to you 726 00:32:20.960 --> 00:32:23.800 by our employer ITProTV the e 727 00:32:23.800 --> 00:32:26.760 learning platform for IT professionals. On if you 728 00:32:26.760 --> 00:32:29.590 go to itpro.tv you can 729 00:32:29.600 --> 00:32:32.490 apply the promo code HaskellWeekly30 at check out 730 00:32:32.500 --> 00:32:35.340 to get 30% off the lifetime of your 731 00:32:35.340 --> 00:32:36.050 subscription. 732 00:32:36.540 --> 00:32:39.520 >> Oh, yeah? Well, thank you guys both for being 733 00:32:39.520 --> 00:32:42.360 on the show today. It has been a blast. Thank you all for 734 00:32:42.360 --> 00:32:45.230 listening. And I'm signing us out. Haskell 735 00:32:45.230 --> 00:32:45.760 Weekly. 736 00:32:46.740 --> 00:32:47.160 >> See you 737 00:32:48.340 --> 00:32:49.070 >> Peace.