3 ms·
Management problem is an engineering problem. Management is how you engineer your organization. I don't know how big the organization the author is working at,
by a_c 4y ago
Management problem is an engineering problem. Management is how you engineer your organization.
I don't know how big the organization the author is working at, too big in my opinion, too big in complexity.
What I observe, could be wrong
- the team "writing" the service is different from the team "fixing" the service. It is like letting one guy write code, the other guy compile code, but on a team level. Feedback of quality is lost.
- "Then I got lucky and intercepted a few of the zanier ideas". It is common in tech companies nowadays, but I think is really not ideal, one party for ideation, one part for implementation. This is not making full use of engineers in my opinion. Give me a recipe, I cook for you. Engineer can be both chef and a cook, but people seem to be happy being a cook. Feedback between ideas and implementation is lost.
- "The surprise came later, when the biannual review cycle ". If it is performance review. It is too sparse. Engineers might already be gone and good luck keeping the service up
- "They wanted to give this person some bullshit sub-par rating". Lacking objective measurement. Maybe that's why they run the review biannually. Subjective rating is time consuming. I imagine it having to compile "evidence" from various source, and give "justification".
You always want the feedback be quick and whole. Code compilation fast, test run fast, CI/CD fast, HTTP return fast, usage gathering fast, the feedback about whether your piece of code is useful fast, the feedback about team's performance fast. In engineering, we take for granted that things can happen almost in real-time. But in management, somehow feedback (reports, performance) need to be "compile" X times per Y period. In engineering, we somehow have the thinking that abstract only when the code is repeated the 3rd time. Hence we avoid abstracting the wrong thing. In management, we often form an abstraction (team) based on some vague notion on how things are compartmentalize. Oh yeah let's have 2 frontend and 2 backend on the completely new recommendation feature. That's the team formation. Engineering and management have a lot in common
- jeremyjh 4y agoThere was not another team fixing the bugs. I understood her original team to be some kind of SWAT team that works across the org to triage and pin root cause of high impact issues and outages. That kind of team deals with incidents, but they don’t fix software bugs. How could they possibly do that across many different teams and codebases?
- a_c 4y agoI did't mean they fix bug. They are to notify "compilation error". They are to triage and tell service team to fix it, rollback the service to a previous version, keep system up. They don't pin root cause, there is always rooter cause. The root to them is probably figuring out which service is down, which is not the root. This all come from SRE. From micro service. From practice most company ain't going to need but adopted. Which is my original intention: not sure how large the organisation is, but felt overly complex