Comment by cogman10 | Hacker News Reader

Comment by cogman10 | Hacker News Reader

Parent Full thread

This all, imo, is simply a trust problem primarily from leadership. Leadership does not trust the grunts to do productive work. So in order to make sure productive work is done, they build elaborate systems of cases, reviews, meetings, planning, scheduling, fighting, readjusting when the schedules are invariably missed, and finger pointing. All almost always completely devoid of input from the grunts.

The middle management problem is they are right in the worst place possible. They are removed from the actual work being done so they don't know what it actually takes to do anything and they are blamed for things not accomplished. Further, they are rewarded for every little stupid thing done. It hyperintensities them to do lots of small safe initiatives and vehemently oppose anything with any sort of risk. All while being almost completely disconnected from what actually needs to be done.

This all leads to a culture meant to squash innovation. Middle management isn't rewarded for implementing a grunt's idea, they are rewarded for delivering CEO initiatives. Anything that takes time away from that is seen as waste.

> Leadership does not trust the grunts to do productive work

Not only that, but unfortunately they are usually right. Without oversight the in-house team has high likelihood of building NIH spaghetti, which causes more problems down the line. To avoid the negative outcome leadership needs to be technically competent and resourced, and that's the other side of the coin - usually they don't have the expertise so in a way they also do not trust themselves to lead the project to a positive outcome.

> Without oversight the in-house team has high likelihood of building NIH spaghetti

I think when you see this you need to start digging deeper and questioning why this is happening.

Is it because the "grunts" are genuinely bad at their jobs? If this is the case, then who hired them?

Or is it because they have been conditioned to believe that if they ask for permission to use an outside tool/library/etc, they will be told "no, we don't have the budget for that" or "that has to go through 12 layers of approval" or "great idea! we'll get it into a committee to talk about the best way to implement it and get back to you (in 6-12 months)"?

In other words: Are they building NIH spaghetti not because they lack oversight, but because they have too much, that hampers them from actually doing their damn jobs?

Don’t forget incentive structures that award developers for “org wide impact”. Often the way to do that is to roll your own crap instead of use off the shelf stuff. This is how you wind up with developers creating their own key-value databases or billing systems… it sounds much better to write these from scratch.

I'd like to add that NIH spaghetti sometimes comes from customer relationships rather than from exuberant developers.

The important customer wants just one tiny convenience-feature that fits their use case and it sorta makes sense... Which somehow keeps scope-creeping over time into a cancerous unplanned product which is expensive to support.

With the benefit of hindsight, you realize it's something the customer should have bought directly for themselves from a completely separate and more-qualified vendor. However they either didn't realize what they wanted or they were able to trick you into building an over-specialized product for their use case--for much cheaper than if they had hired a contractor to customize another better offering.

> they are usually right.

This message encapsulates why so many software jobs are terrible. Put your heart and soul into doing your best, earnestly combat NIH and pursue meaningful productivity, and _still_ the culture is such that at many companies, there will never be trust because the prevailing culture is that management is "usually right" that grunts can't be trusted.

I've worked for good bosses that aren't like this, but they're hard to find.

Clean up enough messes left behind by grunts that were trusted by management and you will be cynical too :-)

I still blame the management here for not building a team where quality and foresight matter enough to avoid large messes.

Trust from management isn’t all that is needed for good work to be accomplished, good managers of software development also should not shockingly know good software development.

I swear software and tech related roles are one of the few places where the managers don’t also know how to do the job of their reports.

Concur. It's an unfortunate race to the bottom which is just indicative of the dilution of ability in this godforsaken industry. Why can't a team of grunts manage a basic CRUD web app? That's what's so frustrating: most of these problems aren't even _hard_, and yet devs screw them up anyway.

Signed, a greybeard sick of cleaning up messes.

For what it's worth, I'm actually a senior IC who's cleaned up plenty of messes over decades and who enjoys migrating legacy systems. In my experience, it's the management-level architecture astronauts who don't understand loose coupling who do the most harm. The damage done by low level individuals can be contained if they're working on properly isolated subsystems (which of course requires competent management to set up).

And I'm cynical all right. I used to believe that I could be part of a "we're all in this together" team. But I've realized how rare that is after bad experiences at multiple companies where management sees an antagonistic relationship with engineering as inevitable — because they agree with you that "grunts can't be trusted".

What's difficult is that usually half the team of grunts can be trusted to run their own race, the other half can be completely clueless and need explicit guidance, even if they are decent programmers otherwise.

If you were to allow everyone to innovate freely, some would spend several weeks learning the latest shiny fad and creating things of no or negative value, like using AI to generate commit messages or adding service mesh to your single container kubernetes cluster. The disconnect from business value and writing code for the sake of writing code can be astonishing. Conversely, if you let loose one smart guy in the team, it may seem unfair to the rest of the team.

A good leader is the one who balances the need and utilizes the best of both these sides.

You’d think from this thread that the messes inevitably originate from ”grunts” wandering off the reservation. In my experience the source has often been “innovation” imposed from above, in the form of buzzword-driven development championed by semi-technical leadership enamored of this poisonous “grunts can’t be trusted” ideology.

> This all, imo, is simply a trust problem primarily from leadership. Leadership does not trust the grunts to do productive work.

Which, IME, stems from a deep-seated classism that sees "grunts" today as being essentially no different than assembly-line workers in the Industrial Revolution: you're just a pair of hands who not only doesn't know enough to make changes in the process, you shouldn't even think of it, because it's not your place.

In this worldview, it's managers (and up) who have the education, intelligence, and breeding to know best, and lowly workers just need to shut up and do what they're told.

...This attitude is also responsible for a lot of other really destructive problems in the modern world of work.

What bothers me about the rest of this thread is not that "grunts" are seen as fallible, but that management is implicitly infallible. Blaming everything on the "grunts" is irresponsible garbage in an industry where leadership, of necessity, often doesn't have a deep understanding of what's going on and the ability to make generally correct decisions with incomplete information is a critical management skill.

Reply on news.ycombinator.com