text stringlengths 0 2.35k |
|---|
[2383.82 --> 2387.68] out like how much things are costing, it doesn't matter if we're just making up terms that don't |
[2387.68 --> 2391.58] actually come back to reality, that's going to degrade the trust from product and business. |
[2391.58 --> 2394.90] So we have to say, okay, we have an 80% probability of this thing happening. |
[2395.40 --> 2397.14] Okay, did it actually happen? |
[2397.42 --> 2399.72] And how many times did it actually wind up happening? |
[2400.08 --> 2403.06] Just going back and like tracking all of that and tracing all of that. |
[2403.12 --> 2406.28] So then every time the business asks, we can be like, here's, you know, the pages and |
[2406.28 --> 2409.92] pages and pages of like how we've discussed this and how we came to these |
[2409.92 --> 2411.66] numbers, how we came to this probability. |
[2411.66 --> 2415.98] And here's like the historical aspect of that and how much of the time, how much it has occurred |
[2415.98 --> 2420.58] historically so that we're actually able to like give you some basis to believe us and |
[2420.58 --> 2422.14] kind of build that trust back up. |
[2422.14 --> 2425.22] And I think that's how you start making it more quantifiable for people. |
[2425.72 --> 2429.42] But I think to your point of that, like type of engineer that we need to develop, I think |
[2429.42 --> 2433.20] one of the types of engineers we do need to develop are people that are kind of focused |
[2433.20 --> 2437.60] on like going into code bases and doing this type of analysis and figuring out how to |
[2437.60 --> 2441.94] prioritize that debt that we have and come up with like debt consolidation and debt pay |
[2441.94 --> 2446.64] down plans and help us actually determine, hey, we want to declare bankruptcy and build |
[2446.64 --> 2447.24] a new thing. |
[2447.64 --> 2448.94] This is how we're going to do it. |
[2449.00 --> 2452.44] So we don't wind up with the old thing still being there haunting us forever. |
[2452.94 --> 2455.22] That is definitely an interesting thought. |
[2455.34 --> 2459.22] But it is like you mentioned, that is almost like a new type of engineer, a new type of position |
[2459.22 --> 2465.58] because coming up with all those percentages and whatnot, that's like an FTE almost. |
[2465.76 --> 2470.72] So it's not something that you can just kind of do on the side, you know, while you're doing |
[2470.72 --> 2471.52] your feature work. |
[2471.76 --> 2474.66] You need somebody who's focused and dedicated to that. |
[2475.22 --> 2475.94] Yeah, yeah, absolutely. |
[2476.14 --> 2480.90] It's like it's a it's a thing you have to be dedicated to doing and actually want to push |
[2480.90 --> 2482.90] your organization forward with doing it in the future. |
[2483.26 --> 2486.64] And it's like a whole process and a whole framework you have to develop as an engineering team. |
[2486.64 --> 2488.86] Like we're lacking this in most of our engineering organizations. |
[2488.94 --> 2494.00] Like I have not been to, I think, any companies that have had this type of robust way of talking |
[2494.00 --> 2497.70] about technical debt or legacy systems or even just feature work. |
[2497.70 --> 2500.16] Like most of the time, it's just like we got a bunch of story points. |
[2500.22 --> 2503.54] We got a bunch of sprints and we're going to take a guess and we're going to have retrospectives. |
[2503.54 --> 2507.68] But retrospectives are complaining about things that went wrong, not about did we actually |
[2507.68 --> 2509.12] meet what we said we were going to meet? |
[2509.56 --> 2513.00] So I think there's definitely a lot of area here that we could kind of build up something |
[2513.00 --> 2515.22] better for being able to discuss a lot of this stuff. |
[2515.22 --> 2519.52] Well, yeah, no wonder if that also calls out there's oftentimes so much ceremony involved |
[2519.52 --> 2523.76] that at least my experience has been a lot of engineers are like, well, this is kind of |
[2523.76 --> 2528.50] killing productive time because like, you know, if we spend a bunch of time pointing things |
[2528.50 --> 2533.82] when we do the work based on scrums, but no one's holding us accountable for our sprints |
[2533.82 --> 2538.06] and they actually end up devolving into something like business ends up wanting con bond anyway. |
[2538.30 --> 2540.48] So I'm doing all this ceremony work for nothing. |
[2540.96 --> 2543.28] Why do I want to add more ceremony onto that? |
[2543.28 --> 2548.44] Yeah, I think that's the kind of hole we've dug for ourselves in building a system where |
[2548.44 --> 2552.60] we are supposed to be getting these benefits, but then not actually doing the back half of |
[2552.60 --> 2553.12] that work, right? |
[2553.24 --> 2557.84] Not actually figuring out like, are these systems like is this form of agile we're doing? |
[2557.92 --> 2559.08] Are sprints working for us? |
[2559.12 --> 2560.40] Are story points working for us? |
[2560.58 --> 2564.76] Are we able to actually get the benefits out of them that we that we think we're getting? |
[2564.76 --> 2567.12] Or is this just kind of in reality? |
[2567.24 --> 2572.12] Yeah, like you're saying a ceremony, something we just do because we've done it, not because we have |
[2572.12 --> 2575.12] we're getting any intrinsic value out of it at the end of the day. |
[2575.32 --> 2579.26] I feel a whole go time episode coming on on this exact topic. |
[2580.38 --> 2584.36] So I kind of I'm interested to hear a little bit more about we've talked about getting to that |
[2584.36 --> 2589.64] explosion point where it's falling apart, you know, having bugs every day, you know, |
[2589.64 --> 2594.36] making your on call engineers life hell, that is the point at which you're like, we need to do |
[2594.36 --> 2594.78] something. |
[2595.12 --> 2600.18] But going back to our original question, like, how do we stop legacy code creeping in as we're |
[2600.18 --> 2602.62] talking about accumulating tech debt, etc? |
[2603.18 --> 2608.86] What are some ways that we can stop getting to that kind of final point of explosion where |
[2608.86 --> 2610.44] everything's falling to pieces? |
[2610.96 --> 2616.48] Well, so I think one thing that like, if you want to keep from code becoming legacy, I think |
[2616.48 --> 2619.12] that testing is a helpful tool for it. |
[2619.12 --> 2621.44] It's by no means a solve it. |
[2621.56 --> 2627.28] But you know, if you can define your actual problems that you're trying to solve and write |
[2627.28 --> 2633.20] tests that exercise those problems, that kind of it'll also help, you know, ensure that |
[2633.20 --> 2638.26] you've got discrete chunks doing those things rather than tangled mess all playing together. |
[2638.62 --> 2641.10] Sounds like you're advocating for test driven development. |
[2642.30 --> 2644.02] I just might be. |
[2645.48 --> 2646.56] It's not the cure all. |
[2646.56 --> 2649.88] There's oftentimes a lot of places where test driven development doesn't work, especially |
[2649.88 --> 2652.74] like lab settings or where you don't know what you're going to do. |
[2653.12 --> 2658.64] But it is a tool that helps in more than just like ensuring that your code functions correctly. |
[2658.64 --> 2665.58] In my experience, you have to invest in your infrastructure and your monitoring probably before you invest |
[2665.58 --> 2670.70] in your in actually rewriting your code in the sense that, for example, if your deployment |
[2670.70 --> 2676.38] pipeline takes hours for your changes to get deployed, you need to rewrite the deployment |
[2676.38 --> 2678.02] pipeline first, right? |
[2678.44 --> 2683.50] Add tests, add monitoring to make sure that you know what your system is doing, like what |
[2683.50 --> 2687.88] queries it's executing on the back end or what external calls it's making or whatever it is. |
[2687.88 --> 2693.82] And probably do that before you get to that explosion point that Angelica just mentioned. |
[2694.06 --> 2696.94] And then the explosion point will be a lot less painful. |
[2697.36 --> 2698.28] Yeah, I like that. |
[2698.48 --> 2705.56] I like that going small piece by piece and just being in a better way each week probably |
[2705.56 --> 2707.02] than you were last week. |
[2707.18 --> 2709.20] It feels like a great, great way to do that. |
[2709.20 --> 2714.30] And also probably less careful and dangerous, if I can say that. |
[2714.56 --> 2720.12] Because when you think about that, you know, rewriting as an entire software is, it is scary, |
[2720.70 --> 2721.14] for sure. |
[2721.78 --> 2725.84] I think I would say writing documentation or just comments, really. |
[2726.28 --> 2728.10] Like I think testing is definitely helpful. |
[2728.10 --> 2734.42] But I think if I were to be given a code base that has really good tests, but no comments |
[2734.42 --> 2738.00] and documentation or a code base that has documentation comments, but no tests, I would |
[2738.00 --> 2738.86] definitely take the latter. |
[2739.48 --> 2744.06] Because good comments and documentation tell me the intent of what this thing is supposed |
[2744.06 --> 2747.06] to be doing, which then I can go write the test for myself. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.