As I grew more experienced at my job, I submitted fewer and fewer CLs, with fewer lines. Eventually, I was just making config changes. You probably would have thought I wasn't very effective but the reality is the small number of changes I made were in exactly the right place at exactly the right time. I've seen other folks who are commit fiends and a lot of their code just didn't stand the test of time, or lacked good tests, or failed at scale, or were just copies of something elsewhere with the class names changed.

I reserve my greatest praise for those who delete code without visible regressions or increased complexity.

I really enjoyed the read.

Thanks!

Personally, I agree with you. I work somewhat differently, but am highly effective.

I managed a team of very high-functioning C++ programmers, for one of the top imaging companies in the world, for 25 years. I kept employees for decades, and that’s no secret. I mention it frequently, here, and my entire career is an open book. I've been an active software engineer for over 35 years, and shipping software, that entire time.

There's a [vanishingly small, I know] chance that I used a couple more techniques than simply counting CIs.

The dao of programming right here

> Just look at the commit logs.

Yeah no. I've had to do a lot of investigative, hacky "PoC" work with lots of debugging, design, coordinating work (whiteboarding, email threads, etc) for the last ~5 or so months. I'm a software engineer but if you looked at my actual commit logs for the last quarter or so, it'd seem like I haven't done much. Some people work differently and have different workload, I've been shifting to some more active work now (and I'm glad because I enjoy coding way more) and my commit log "landscape" will probably be very different in the next few months, but if you just looked at the last few months it'd look like I didn't work a single day.

I really hope people don't base productivity just on commit logs, seriously.

I’m always amazed at how quickly we jump to judgement (and finding wanting).

I was a very good manager for 25 years.

I guess maybe I ended up doing OK. Read the testimonials on my LikedIn.

I’m pretty aware that CIs are simply one single data point.

Some of the days with the fewest commits, are the ones I did the most work.

That is an extremely poor way to view the work output of a SWE. I sincerely hope that is not how you actually determined productivity for your team.