[dewey]

How actionable will these insights be? Are you going to write better code because you don't want your "failed test" metric to go down?

As with many of these "quantified self" stats it feels like it will result in a colorful and nice to look at dashboard...with no benefit.

[hirako2000]

These "quantified self" stats provide countless insights.

One example.

There is a refactoring view in there. By time, by projects, even breaking the figures down by type of actions.

Benefits:

- Self awareness. It is hard to gauge how much time is spent refactoring. If your priority isn't to refactor but meet a soon coming deadline, these stats tell you whether you are adhering to the priorities.

- Quantifies. If you are trying to explain to your colleagues that you find yourself needing to do a lot of refactor for that particular project, you've got numbers to communicate. What's a lot? Some colleagues often ask.

- Evidence. Showing these numbers communicates better certainty than "I think I've been doing a lot of refactor on this project today"

Plus, oftentimes with visualisations, we don't know what we are looking for. Until we find it.

[The_Colonel]

I use refactoring tools all the time during development and bugfixing. The distinction between refactoring / feature development / bugfixing is mostly in the intention. If it just tracks the usage of refactoring tools, I think there will be many false positives.

[emhauck]

Yes, metrics alone are not sufficient here. But I did not want to include opinionated targets where optimization towards might be problematic.

For me, I had 2 reasons to look at the metrics: First, I wanted to split my work into smaller chunks and commit more often, and I wanted to track whether I achieved this goal. Second, it occurred to me that I was using some IDE refactorings a lot of times, but I was wondering where my „blind spots“ where, i.e which types of refactorings I was using rarely or wasn’t even aware of. This inspired me to track IDE refactorings in the plugin as well.

However, there might still be the use case of some kind of „trainers“ that could be included in the plugin that help you improving your coding. Running tests more often, continuously integrating your branches, committing more often - these are not silver bullet mechanisms, but they do make sense a lot of times and a tool might help here.

[cratermoon]

It's data. Creating a colorful dashboard turns it into information, To get knowledge and action is another step. The plug-in does what it says on the tin, it's up to the user to make something of it, or leave it as inert knowledge.

[Ragnarork]

> How actionable will these insights be?

I mean, that's something you should try to answer yourself, if you think you can extract any benefit from having these stats.

To me it's similar to a fitness app that can tell you things like time, speed, distance, elevation, but won't really tell you how to run better.

I have a hard time coming up with ways this could help my coding habits right away, but I think this would be on the user to find these, not the stats reporting tool, no? And if you find no real use, then it's maybe just not for you.

[The_Colonel]

> To me it's similar to a fitness app that can tell you things like time, speed, distance, elevation, but won't really tell you how to run better.

Most of these stats are very simple to interpret - higher speed/distance - you're getting better.

My fear is that people will apply such simplistic evaluation on these stats as well - "your daily commit rate has been going down lately, we need to focus on that".

[latchkey]

It'll tell your boss if they should keep you around or replace you with AI.

[xyst]

1. Create local (or private remote repos) with similar name as remote repos in VCS

2. Use scripts to commit junk to local or remote repos

2a. Extra points if you use chatgpt, Claude, gemini beta alpha0 to generate junk commits

3. ???

4. Profit. Sit back for a few months or quarters. Interview for new jobs, and then bounce out of there