text stringlengths 0 1.82k |
|---|
**Carlisia Thompson:** Yeah. |
**Brian Ketelsen:** But how many times do we catch ourselves not reading the error? I do it at least once a week. Or the error message is staring me right in the face, telling me exactly what's wrong, and I change 16 other things trying to fix it before I realize what I've done. |
**Katrina Owen:** Yeah, I'm always making assumptions. |
**Brian Ketelsen:** And then I feel foolish. |
**Katrina Owen:** Yeah, because we are all foolish; let's just be honest here. |
**Carlisia Thompson:** I think we are in a hurry more than anything else. |
**Katrina Owen:** Yeah, I think we are. About six years ago I was freaking out basically, because I thought I had to learn everything. Like, every new article that I saw was another thing I had to learn, and I had this backlog of hundreds and hundreds and hundreds of articles, tutorials and technologies I had to learn. |
And I finally decided that I was gonna give up. I was not gonna learn anything. I was gonna only do research or learn something new for one of two reasons. The first one was if I had an error, I would slow down and figure out exactly why I have the error. No more jumping on Stack Overflow and guess, like copy and paste... |
**Brian Ketelsen:** Yeah, that's a good rule to live by. |
**Erik St. Martin:** Yeah... I need to go through and clear out all my bookmarks of things that I wanna learn, and accept that some of the stacks of books that I bought I'm just never gonna get to. Because I tend to want to dive in; there'll be some new thing that I'm challenged with, and I'd be like "Oh, that seems co... |
**Katrina Owen:** Yeah. I try to put off actually bookmarking or searching or buying until I sit down and decide to do it, because it's like, I know how to google, I can find those links if I want to; I can find the books if I'm ready. If I have a whole weekend to spend on something, I'm sure I'll find good resources. ... |
**Erik St. Martin:** That's true. You just kind of do it real time, rather than pre-plan your learning flow. |
**Katrina Owen:** It's hard though, because it's tempting. |
**Brian Ketelsen:** Speaking of errors, our other sponsor is Backtrace. |
**Break:** \[39:28\] |
**Erik St. Martin:** That was perfect. If felt kind of right in. |
**Carlisia Thompson:** We planned it. It was all planned. Everything we do is planned. |
**Erik St. Martin:** We all knew the way the topic was gonna flow. So I think we've got about 20 minutes left, and I wanted to get to a technology you had mentioned in the email, Katrina, which is the GraphQL API. That's been something that has been on my radar for probably a year or two, the whole notion of GraphQL. I... |
**Katrina Owen:** Let's do it. Should I start? |
**Erik St. Martin:** Yeah. |
**Katrina Owen:** I was gonna write a little tool a couple weeks ago, and this was a tool that the idea is to use the GitHub APIs to get more data about the health of a repo or the health of a project. By health, the things that I'm thinking about the most are how responsive are the maintainers. If you post an issue or... |
The other thing is, are there any people actually commenting on pull requests and issues? It doesn't even need to be the maintainers. If someone's giving code reviews, it doesn't really need to be someone who has a commit on the project. So I wanted to use the GitHub APIs to find this information, because it's not obvi... |
I realized that I was gonna have to make basically a bajillion requests, and for every request that I made, I was gonna have to make an extra request to figure out some ancillary information that I needed in order to do this analysis. This was like a week before GitHub announced the GraphQL integration. |
So the thing about GraphQL is you can now design your query upfront and get all of the data back and not have this N+1 problem when talking to the API. That's pretty exciting. |
**Erik St. Martin:** I didn't even realize that GitHub had GraphQL APIs now. The only place I had really seen leveraged was through Facebook where it came out of, but it's really cool, the way you nest the information that you want, and then kind of have these little -- I don't know what you would call the little addit... |
It was really cool thinking about that, where you have this highly nested data where you don't wanna have your looping problems, and stuff like that. So have you actually had a chance to work with GraphQL? |
**Katrina Owen:** No, it's on my list for when I get home from this trip. |
**Erik St. Martin:** \[43:45\] I have to admit I have yet to work with it, but it seemed interesting. A couple of years ago Brian and I were working on a project that had that kind of nested data where you wanted to get this tree of information and do counts on some of them, and things like that. We ended up designing ... |
**Brian Ketelsen:** No. |
**Erik St. Martin:** It was something that Facebook had started talking about. |
**Brian Ketelsen:** It was unreleased at that point. It probably would have worked, but maybe it wouldn't have been fast enough. I don't know. |
**Erik St. Martin:** So in your concept of doing this intelligent querying against the GitHub API, it's just to kind of rate the contributors to a different project to kind of determine who should have what role, or...? |
**Katrina Owen:** Yeah, there are two things. One is where do I need to put my attention? I might have 70 repositories, and it's really hard to know what the state of everything is, and I get 500 notifications every week about issues that I need to look at. So I need to know if there are repositories where the maintain... |
Then for the ones that don't have responsive maintainers, I can spend more time. I can filter that into something that I check every day and make sure that I give the feedback that needs to be given on a regular basis, and much more aggressively. The other thing is for a repository where someone is giving feedback and ... |
I had a working prototype where I was sucking down the data, but I was using the old APIs, not the GraphQL APIs, so it wouldn't work to get me all the..It wouldn't scale. |
**Carlisia Thompson:** Besides the use cases that you just gave us for the maintainer, for the consumer of the project, I think it would be super awesome. Every time I run into a project that I'm considering, unless it's a super well-established and well-known project, there is a checklist of things that I do. I go thr... |
**Katrina Owen:** Yeah, this would potentially... What you said - run it against the repo and say, "Yes, people are getting responses." That would actually work even if there's not a lot of activity. If people get a response very quickly, even though the project is mostly stable, then it's still a good project to contr... |
**Carlisia Thompson:** Super cool. |
**Erik St. Martin:** Yeah, that's awesome. Although I have to say I've never worked on a project so big that it would require hitting the API to determine the state of my project. That's a whole new world. |
**Carlisia Thompson:** But sometimes you might not be a maintainer that works on a project that's so big, but you might have a ton of little projects, and it takes a lot of time to be on top of everything. |
**Katrina Owen:** \[48:02\] Yes. |
**Erik St. Martin:** Doing open source stuff on the side and also holding down a day job, it's a struggle for a lot of people, so anything that can kind of help organize things and tell them where their minimum amount of time is best spent, I definitely think it's advantageous. |
**Katrina Owen:** There's a really good question in the chat right now: what if the project is run by someone who closes the issue instantly and dismisses them? That's really not healthy behavior at all, and I don't know if we could surface that with the data. Are they also giving a response when they close it? If all ... |
**Erik St. Martin:** Yeah, because I think there's a lot of factors that go into that too, right? Maybe they're quick fixes that require commits, maybe it's a duplicate of other issues... There's probably somebody who's better at math than I that can figure out some insights into whether it's unhealthy closing of issue... |
**Katrina Owen:** Yeah, so thanks for the question, Florian; the question was mentioned in the GoTime FM Slack channel. |
**Brian Ketelsen:** Yeah, how awesome is it that our guests are answering questions live on Slack while we broadcast? This is 2016, folks. This is it. This is 2016. We have reached the future. |
**Carlisia Thompson:** Katrina, you also have a blog that you post to... Maybe not super frequently, but the content is amazing. I know you have one blog post that coincidentally last week Aaron Schlesinger took one of the concurrency examples... Three Trivial Concurrency Exercises For The Confused Newbie Gopher. He to... |
**Katrina Owen:** That's awesome, I didn't know that. |
**Carlisia Thompson:** Yeah, I'll link to it. It's super helpful. |
**Katrina Owen:** That's great to know. I just kind of write... Sometimes I'll run into a question that it seems like there should be a good answer, but there is no blog post that really does that, so I try to just write that blog post that somebody might stumble on and could be useful to them. The most recent thing I ... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.