text stringlengths 0 2.35k |
|---|
[817.44 --> 821.38] But I think there's ways in which even those types of code bases can be a pain to work with. |
[821.38 --> 830.32] But I would absolutely say like a code base that has zero tests, no documentation, none of that is like the worst kind of legacy code bases that we can deal with. |
[830.32 --> 835.58] How do you know then if your test suite is going to be good for the long run? |
[836.08 --> 838.02] Good design of your software. |
[838.72 --> 844.60] Like I think like if you kind of got to push the boulder up the hill or like kind of really go back to the genesis of the stream, right? |
[844.60 --> 854.16] It's a lot easier to like if you're going to build a dam, it's a lot easier to build a dam across a stream that's two feet wide than it is to build one across one that's 200 or 2,000 feet wide. |
[854.38 --> 862.02] So, yeah, it's like if we put a lot of the stuff downstream to know if we've done the right thing, we can through various techniques understand how good it is. |
[862.02 --> 868.08] But I think it's much better to try and put in ways of knowing how correct your code is upstream. |
[868.30 --> 872.92] But of course, too, it's like a stream, a dam across two feet of stream is quite useless of a dam. |
[873.24 --> 876.50] So there's a middle ground here of where we need to find the right place to put things. |
[876.54 --> 883.38] And I've definitely found that right place to be in good design documentation and good design ideology for the software that we build. |
[883.74 --> 886.76] But I'm sure other people have different opinions on that. |
[886.76 --> 891.00] And I think I'd add modern code base does also does not mean tested. |
[891.46 --> 897.66] There's plenty of fresh code that I see written that like engineers oftentimes don't like to write tests. |
[897.96 --> 904.88] So that is a pervasive problem that seems to plague all code bases and stuff just isn't tested. |
[904.88 --> 916.16] Yeah, one thing I can say we've done to kind of mitigate that problem is to keep an eye on the coverage thresholds and actually building the test coverage thresholds |
[916.16 --> 918.68] into our testing and deployment pipeline. |
[919.12 --> 924.24] So if someone adds new code without adding tests, the pipeline actually fails. |
[925.02 --> 927.30] So that is definitely an interesting approach. |
[927.36 --> 930.06] But I've also seen it backfire horribly in the past. |
[930.06 --> 941.00] I've seen engineers and this actually got somebody talking to, but like this is in Java land, put an entire class on a single line because they were just annoyed by the test coverage requirements. |
[941.00 --> 946.18] And so you can kind of get malicious compliance with those code coverage requirements. |
[946.38 --> 949.60] And also depending on your language, like Go, it's not as bad. |
[949.70 --> 956.12] But in Java land, there's lots of all sorts of like you've got to catch these exceptions that practically are never going to happen. |
[956.12 --> 959.46] But you might not even be able to write tests to get there. |
[959.88 --> 963.32] So I think you do have to be a little bit careful with coverage requirements. |
[963.54 --> 965.66] I think it's a good metric to have and to track. |
[965.98 --> 969.98] But there's a danger zone when you start failing builds because of it. |
[969.98 --> 974.54] I mean, I'm interested to hear a little bit more about your reference kind of Java, Go. |
[974.54 --> 987.86] Are there aspects of Go as a language that make it easier for you to either avoid quote unquote legacy code or avoid it being a whole huge refactor project? |
[988.04 --> 993.46] Because in fact, refactoring, you know, bit by bit, you know, backwards compatibility comes to mind. |
[993.64 --> 998.44] Are there aspects of Go that make it easier for us to avoid legacy creeping in? |
[998.44 --> 1002.26] So I think there's some in that it's a simpler language. |
[1002.38 --> 1004.20] Up until recently, it didn't have generics. |
[1004.28 --> 1010.96] So you're not going to have people going nuts with generics and creating really complex things like trying to deduplicate everything through generics. |
[1011.20 --> 1012.12] So that helps. |
[1012.46 --> 1019.44] But I think the biggest thing that Go has going for it with like the lack of legacy and maintainability is that it's a newer language. |
[1019.44 --> 1021.30] It hasn't been around for that long. |
[1021.68 --> 1029.86] So you don't have projects that have been sitting under or that, you know, have been that were conceptualized 20 years ago to carry that forward. |
[1029.98 --> 1034.66] But there's a lot of Java projects out there that have been around for a long period of time. |
[1035.08 --> 1041.62] And so I think there might be a false sense of security saying, well, Go is going to have less legacy just because, you know, you don't see it as much. |
[1041.62 --> 1048.62] But that's also a function of the language being newer and therefore not having, you know, just the crap that's been around for a long time. |
[1048.98 --> 1052.66] And that's also something like we learn every time we write code and every time we develop. |
[1053.02 --> 1060.46] So there's a lot of patterns that, you know, we might have been doing 20 years ago in Java land that, you know, we've maybe learned better ways for that are being let go. |
[1060.86 --> 1065.46] But again, that's not so much a function of the language, just of the newness of it. |
[1065.46 --> 1082.32] I just wanted to add that what John was saying kind of made me think about what this topic of legacy will be like for Go in like 10 or 20 years in terms of when it gets to be old enough to like the same age as Java is currently. |
[1082.68 --> 1087.04] And we have a lot more Go projects around that need maintenance. |
[1087.26 --> 1090.90] It's just an interesting thought experiment to be having. |
[1090.90 --> 1097.96] Well, I can tell you about stories that, you know, with .NET, even though, let's say it started in 2001. |
[1098.16 --> 1105.24] So in 2005, there were already projects that you could not even migrate to the next version. |
[1105.62 --> 1113.68] So to me, the tooling that the language provide, which is very solid in Go, in my opinion, and the fact that it's backward compatible, |
[1113.68 --> 1119.48] if they can maintain that for a long time, it will diminish the effect of that. |
[1119.48 --> 1120.56] Yes, I'm certain of that. |
[1120.64 --> 1126.82] Because in my case, at the moment, the tooling of Microsoft is my main issues. |
[1127.20 --> 1128.26] It's not the code per se. |
[1128.44 --> 1131.02] I mean, it's not that at all, but the tooling. |
[1131.64 --> 1136.16] And I guess I'd be interested that because I've never really worked in .NET, always been allergic to Windows. |
[1136.16 --> 1146.14] But the Java side of things, you do have, like, you know, but they strive for a long time to make the JVM backwards compatible with prior versions of Java. |
[1146.38 --> 1154.66] And that didn't necessarily stop, like, the need to have complete rewrites or, like, you being stuck on an ancient version of Java because, like, whatever new thing. |
[1154.66 --> 1165.86] I've actually even seen the new thing with Go where they're seeing projects that were, like, running on old versions of GAE that have pinned their version of Go to, like, 111 or something. |
[1166.10 --> 1168.20] And, like, upgrading those has become a nightmare. |
[1168.20 --> 1171.76] So, again, it's less likely in Go. |
[1171.88 --> 1178.56] And it is, for sure, easier to generally upgrade versions of Go than my experience has been with Java. |
[1179.00 --> 1187.58] But, again, in Java land, you're also dealing with a ton more dependencies because there's this massive ecosystem that abounds where you're not having to do everything. |
[1187.78 --> 1191.08] Where in Go, a lot of that ecosystem doesn't exist. |
[1191.24 --> 1195.04] So you're reimplementing a lot of things and not sucking in as many dependencies. |
[1195.04 --> 1201.26] So the benefit of that is when you continue to upgrade, you're only worried about your code, not, like, a million different libraries. |
[1201.96 --> 1208.96] I'd be curious to see if that easy to upgrade continues as the Go ecosystem grows and as it becomes more complete. |
[1209.28 --> 1211.32] Yeah, but you are talking about the platform. |
[1211.48 --> 1213.96] So the Google App Engine is a platform. |
[1214.14 --> 1218.90] It's not really Go's fault or responsibility, in a sense. |
[1218.90 --> 1225.40] You can still take code bases from, like, early, early days of Go and compile that with the new compiler. |
[1225.60 --> 1228.56] This is something that you just cannot do in the .NET world. |
[1228.76 --> 1231.16] So that's mostly what I'm talking about. |
[1231.30 --> 1242.94] When the tooling is failing you, it's a different ballgame because you have the code, but you have the technologies that were suggested to use. |
[1242.94 --> 1255.26] When a company makes a choice of going with something and three years later, it's not even supported by Microsoft anymore, then it's a different thing than bad design, I think. |
[1255.60 --> 1263.54] Yeah, and I guess there's a different user or past user story there and experience because, like, again, coming from Java land, like, everything will compile just fine. |
[1263.54 --> 1271.68] But the things that I've encountered that I've seen that make, like, upgrading hard isn't so much the compilation, which it sounds like in .NET that might be a thing. |
[1272.04 --> 1276.78] But it is the platform that is the hard thing to upgrade from my past experience. |
[1277.20 --> 1288.98] And, you know, so that's where I think that Go won't be immune to the platform increasing scope and, like, hardness to upgrade with that as it becomes more robust. |
[1289.68 --> 1292.34] Or there's more things going to be, like, GAE and whatnot. |
[1292.34 --> 1295.04] So I think I definitely agree with you. |
[1295.12 --> 1299.32] Yeah, if your vendor is making stuff that doesn't compile it in three years, oh, my God, that stuff's terrible. |
[1299.88 --> 1300.38] I'm sorry. |
[1301.04 --> 1302.38] Yeah, I can give another example. |
[1302.38 --> 1308.00] So, you know, our TLS 1.1 just vanished recently. |
[1308.22 --> 1316.88] But there is still an old application that runs under an old version of Windows Server 2008, which is not supported anymore. |
[1316.88 --> 1323.54] But because their migration path to the next version of .NET is not an easy one. |
[1323.80 --> 1325.30] I'm not saying it's not possible. |
[1325.54 --> 1326.36] It's not what I'm saying. |
[1326.46 --> 1331.64] But there is a lot of investment that the company needs to do to migrate to the next version. |
[1332.30 --> 1335.30] So that's another aspect of legacy. |
[1335.30 --> 1344.30] I mean, when the OS and the bare-bone communication system, the protocol starts to fail you, what do you do? |
[1344.38 --> 1355.20] So I had to start writing some very, very small piece in Go that just do HTTP call because I needed to use TLS 1.2 or 1.3. |
[1355.20 --> 1358.18] So, yeah, it's a different ballgame, I think. |
[1358.64 --> 1366.30] That does bring up, like, is working in legacy with closed-source systems different than working in legacy with open-source systems when it comes to those dependencies? |
[1366.72 --> 1368.28] Because that is one thing. |
[1368.32 --> 1372.02] That's one thing I've always liked about working in all the various different texts. |
[1372.20 --> 1373.54] I've been in its open-source scene. |
[1374.90 --> 1376.16] Oracle buys Java. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.