text
stringlengths
0
1.49k
I think that was also a sticking point that really got people setting the new bar. If you wanna make a project that people can try out fast, it needs to be almost a single download, and off it works.
**Erik St. Martin:** \[11:47\] Yeah, that's a fair point too, because everything before... You had conflicting Ruby versions or Python versions. Even in dynamically-linked binaries you can get into dependency issues. So I think I'd agree that just having that single binary, like "Hey, here's the 64-bit Linux version of...
**Kelsey Hightower:** I wanna switch gears, I'm actually stalking Mat on the internet right now, and his GitHub repository is a goldmine. There's these things like bitbars, running scripts inside of your MacOS bar, Gopherize Me... That's a pretty epic project, I see everyone making their avatar using that thing. And yo...
**Mat Ryer:** Yeah, I'm a terrible businessman... \[laughter\] That's one thing that I've learned. So I don't really know... I think if you have to do something to solve the problem for yourself, there is just something very nice about the idea that you've just solved that problem for other people as well. So that's wh...
Bitbar was an interesting experience... That's actually an Objective-C app and as you said, it puts the standard output from any executable, and it takes each like and puts it into the menu bar on your Mac. It's a dead simple idea, it was something I needed \[unintelligible 00:13:17.00\] and I wanted to see it in my me...
If you look through the getbitbar.com, there's a wealth of different things in there that you can just put into your Mac menu bar, and it's all free because it's actually not that hard to write. It's easy to put stuff in the menu bar. The nice thing was the idea of abstracting it and letting people just write scripts t...
**Kelsey Hightower:** Mat, I've gotta ask... Do you keep stats, do you check the star count on your repository? The reason why -- I have this theory that people that do serial repositories (I'm guilty of this, too)... But do you check the stats? Do you look at the star counts? Do you know roughly the number of stars yo...
**Mat Ryer:** I do know on some of them, because that is a way for me to gauge interest. You also know anyway by the interactions you have in issues or pull requests, and things... But knowing which projects are taking off or getting interest I think is a nice way to know where you should put your efforts. A lot of tim...
**Erik St. Martin:** I think I'm a terrible businessperson from the perspective of I build stuff, I put it out there, and then I forget about it. I don't even look. Two years later I'll go to pull up the project and there's pull requests, there's issues open, and I would just wanna be like, "I'm so sorry..." \[laughter...
**Kelsey Hightower:** Yeah, Mat, you make up a good point. I think that's the perfect use case for the stars. I look at them as like thank-yous and true indicators, because it seems like most people rarely give out a star unless they really are either using the project or they really appreciate the content that they fi...
\[15:57\] On Twitter, people just put a heart on everything, like "Hey, I fell down." It's like, "Really? Are you enjoying my pain, or did you really find that to be an incredible tweet?" But GitHub stars seem to have a little bit more substance to it, especially when you look at who's giving you the star.
**Mat Ryer:** Yes.
**Carlisia Thompson:** I think so, too. At least for me, I only star things that I would actually like to find again, even if I don't remember I starred it. It's like a bookmark for me.
**Erik St. Martin:** That's interesting, the way you use the things on the different platforms are different. You're right, Twitter is more like a nod, like "I saw this. I'm listening, I hear you." It's not like, "Oh my god, I love this." Most people just use it as a way to recognize that they're listening to you, but ...
I don't go through an unlike somebody's tweet, but I will unstar a GitHub project.
**Kelsey Hightower:** Mat, tell me about this Gopherize Me. I've been seeing everyone, or at least a lot of folks in the Go community, with these little gopher avatars, and I'm like, "Where are people getting these things from?" and then I discover this thing... I admit, I spent about 30 minutes trying to get my gopher...
**Mat Ryer:** Yeah, Gopherize Me was a very interesting... The way it happened was kind of crazy. The same way you saw these little cartoons appearing, I saw a couple of them - Mark Bates got one, Brian got one, Erik got one... You got one, right?
**Erik St. Martin:** Yeah, so Ashley McNamara did all the artwork for the new GopherCon site, and when she did that, Brian was talking to her and she made him an avatar, and I'm like, "I want a cool avatar!", so Ashley made one for me. Then, of course, Mark Bates was like, "I want one, too!" and that's kind of what lea...
**Kelsey Hightower:** Hold on, so let's give a good shoutout to Ashley, because she has been doing a fantastic job with all this artwork, kind of an extension of all the other artwork we saw in the community... But man, she's been so consistent with it!
**Erik St. Martin:** Every time somebody makes a request about some trait that they'd like to see, she's like "I'm on it. I got it!" Mat took the lead and had some reference code and stuff that was like, "I can turn this into something people can do themselves."
**Mat Ryer:** Yeah, somebody on Twitter said, "You should build a site that just lets people build their own", and I actually replied and said, "I'd love to do that." I was at a point where I needed -- because I work at Grey Matter, and we work on a big project that's a long-running project, and sometimes you need a br...
So this was a perfect timing thing... It was one evening, I had an evening to spare, and I thought "I can quickly put this together." I'd never met Ashley, but jut over Twitter I said, "Let's put some of your artwork into this Google Cloud storage. I'm gonna see if I can load it through an AppEngine app, render it and ...
Then she put the assets in there and we then figured out a few little rules, like all the assets should be the same size, and then we'll just rely on the fact that they are -- we'll layer them up in a certain order in order to build up the picture. And it kind of just happened very quickly. It was less than five hours ...
\[20:15\] It was actually quite slow. Also, the thumbnails originally were the full-size images, because generating thumbnails is something you'd wanna do to optimize it, but initially it works without doing that. So it was very MVP, very much "whatever we have to do for the absolute minimum to make it work", so that p...
I actually shared it before I was even ready, but it was kind of working. Then it started to get traction. It's had about 15,000 users. I put a Google analytics on it from the very first version, because I was interested if it was gonna take off. Suddenly then I started noticing things in AppEngine like "Okay, there's ...
**Kelsey Hightower:** Well, fantastic job. I think there was another important thing that you said in all that - getting those quick wins. I've seen a lot of people - even who don't use Golang as their primary language - use it as their escape language. When they go home, they just wanna build something and get it done...
**Mat Ryer:** Yes, and I used to do it in Rails, too. The only difference is now, especially when I build it for AppEngine, I know that if I put it there, it is ready to scale - it REALLY is ready to scale. I'm confident, as long as there's budget for it, if a million people suddenly went and started gopherizing themse...
Go, the default HTTP package, the comprehensiveness of that - that is something that I think we sort of take for granted a little bit. We just write a handler, it's easy, and we can just bind it to a path and that's it, we're done. But really, that is a highly concurrent web server going on there. Especially in AppEngi...
I like the fact that you can be rapid with Go, but also it really is ready to scale if a project takes off. I'm convinced if this was a Rails app and I just had it on a little server somewhere, I would have seen many more problems with the traffic it got. It did on the first day get thousands of users hitting it straig...
**Erik St. Martin:** So challenge accepted! We need a million people to gopherize themselves, to prove it out.
**Kelsey Hightower:** And make sure you ping me offline... Me being at Google, I love this authentic advertisement for Google Cloud right now. This is totally legit, I appreciate it. But we can definitely help you in terms of credits for the community; we definitely don't want you to bare too much of the cost for somet...
**Erik St. Martin:** \[24:21\] Yeah, I absolutely love all the grassroots efforts like that. So now is probably a good time to take our first sponsored break, and then we can kind of jump in and start talking about some other things, other projects you're working on. Our first sponsor for today is Backtrace.
**Break:** \[24:38\]
**Erik St. Martin:** So we are back, we're talking to Mat Ryer. We were just talking about Gopherize Me before the break, and the grassroots effort of that. Have you had any other projects that have really taken off that surprised you like that?
**Mat Ryer:** No. \[laughter\]
**Kelsey Hightower:** I loved that answer. Straight to it.
**Erik St. Martin:** It's amazing when stuff like that does that, though. It's a little thing that you don't think much of at the beginning and then it just blows up into something much bigger than you expected it to be, and it's really gratifying. I love seeing that.
We were joking last week on the show, all the gophers... It was like I couldn't even look at Twitter anymore, it was nothing but people posting the gopherized versions of themselves.
**Mat Ryer:** I know, I loved it too, and it hasn't stopped... It's actually still growing. We also changed -- you could originally only share it on Twitter, and we added different ways of sharing it, and stuff. We've actually noticed that people outside of the Go community now have started making these little cartoons...
And of course, the source code is available, you can go look at it. And honestly, it's not awesome... It sort of needs a rewrite, and my M.O. usually is "I'll hack at something so that I can sort of understand how to build it, and then I'll do a rewrite of it properly", and usually I'll TDD the rewrite.
The first time I write something, it's sort of like learning, it's like a discovery process where I'm figuring out how to do it, and then the second time I write it, I'll actually write it properly, knowing all the things I learned from the first time.
I think some author said that the art of writing is rewriting. I think we're talking about novels and stuff, but it also applies to software - the second time you write something, it's so much better because of everything you've learned.
**Erik St. Martin:** Yeah, you've got a better understanding of the domain, so if you spend just even a little time - 15 minutes - spiking out a prototype, your new approach is gonna be completely different based on what you learned, and I think that's important.
**Mat Ryer:** Yeah, I agree completely. In fact, I think almost the worst time to design a system is at the beginning of a project, because that's when you have the least amount of information about it. Unfortunately, that's when we have to do it, but there's definitely something to be said for jumping in, just getting...
\[28:12\] I genuinely do usually write things twice, and it sounds wasteful, I think, to a lot of people, but the value is there; I'm gonna do it on Gopherize Me. I'll tag the repo as well, so that people can see the before and after. There's bits in there now that can be cleaned away, and some things could be a little...
**Carlisia Thompson:** That's a very good point about rewriting. Sometimes we might have such a big attachment to code, and code is not really the point. I mean, at least that's the way I see things now after having been programming for a while. Code is just something you output in the process of learning the domain, a...
**Mat Ryer:** Yeah, I completely agree. And Go does this for us. Go is a minimalistic language, it's a simple language, so you don't tend to invest too much time; as long as you are building things, discrete little packets of functionality, you can quickly get something that you need, get something that's working. And ...
**Kelsey Hightower:** Yeah, so speaking of improvements, I was going to ask you about Go 1.8. That release is on the way... What are your thoughts about 1.8? Is there anything in there that you're excited about, you've been waiting for for a long time? What are your thoughts on 1.8?
**Mat Ryer:** I'm very tempted to say no again and just leave it, but actually there is defer; defer performance is now improved by a hundred percent, or something like that. It's like half as efficient, or something. Defer is a bit of an interesting subject for me, because I love that you can express the intent in the...
\[31:54\] I've seen people complain about the performance of defer, because there's a performance cost to it, but I honestly think we're obsessed with that kind of performance versus the productivity gains you get by just being able to glance at code that uses defers and understand implicitly what it's doing, rather th...