Build Latency, Predictability, and Developer Productivity | Hacker News Reader

Build Latency, Predictability, and Developer Productivity

computer.org

39 pointskiyanwang12 comments

Seems like the main takeaway for the rest of us is these builds were already absurdly fast and they made them faster. They say “a few seconds” was 13% of the median build time. My C++ builds are significantly slower than 8x a few seconds.

Reading between the lines we can see how the authors struggled to write an interesting paper while walking on egg shells...

Staying "in the flow" is difficult, but build time is clearly not the unique or maybe even the most important factor.

Good tooling should be expected to not break the flow, but from my experience, interruptions and meetings are far worse.

If interruptions and meetings are the worst thing to break your flow then I feel like you haven't dealt with truly bad build systems where builds take multiple minutes to run on a good day.

Build times can be the most important factor if they take like 20 minutes. Which I actually have seen before, I've seen some monster C++ and Java builds in my time.

The developer experience is completely different in such an environment.

As in, productivity goes to shit.

Mostly agreed, but don't discount tooling quality as a major source of flow-breakage.

There's often friction and lack of documentation about tooling. This results in much more frustrating flow-breakage IMO. This is something that we could fix but choose not to because we can't clearly articulate the loss from flow-breakage.

I want to find part 1 & 3 of this series and cannot. Any links would be appreciated.

Part 1: https://www.computer.org/csdl/magazine/so/2023/01/09994260/1...

Part 2: https://www.computer.org/csdl/magazine/so/2023/02/10043615/1...

Part 3: https://www.computer.org/csdl/magazine/so/2023/03/10109339/1...

ty!

I clicked on one of the authors of the paper and it listed all of their work

Part 1: https://www.computer.org/csdl/magazine/so/2023/01/09994260/1... (10.1109/MS.2022.3212165)

Part 2: https://www.computer.org/csdl/magazine/so/2023/02/10043615/1... (10.1109/MS.2022.3229418)

Part 3: https://www.computer.org/csdl/magazine/so/2023/03/10109339/1... (10.1109/MS.2023.3242137)

Having been responsible for running the DOI system for several years, I decided to include DOIs in the parenthesis.

ty!

I love that they explicitly list the xkcd compiling comic in their references.

Reply on news.ycombinator.com