Hacker News Reader

Comment by 100011_100001 | Hacker News Reader

Eh, that is exactly what I don't want. Pretending I am this amazingly productive developer and my code is just trash.

I just checked. In 2021, only two defects were created against my specific code changes. If I had to home into a weakness of mine is that sometimes I solve today's problem well, but don't think extensively about how tomorrow's problem in that space might look like.

The struggle is between creating simple solutions vs over-engineering for future proofing. I'm pretty sure I'm biased towards simple solutions.

View on HN

> The struggle is between creating simple solutions vs over-engineering for future proofing. I'm pretty sure I'm biased towards simple solutions.

I want to encourage you to keep learning about your own biases here but to continue to have that bias as much as possible.

I like the term "throw away code" here but I think another way to say it would be that most of the code I write is like a jig for woodworking. I don't need it at the end of the day, it looks like shit, and holds up the real Things while I figure out how those things actually work.

Sometimes in code though, I find, that the jig _is good enough for now_ and it's super easy to replace the three that break instead of all 20 I needed to build it.

Have you thought about, or are you already involved in interviewing new developers for the company?

It sounds like you might already have a nose for what makes a good dev (i.e. somebody who holds the same values as you) It's a great way to influence the company culture, and get good co-workers, which can also de-stress you as you now have fellow travelers whom you can trust with tasks.

It's also fun, sometimes humbling but always interesting!

It's not about behavioural defects. It's that the _code_ is trash. It is unmaintainable and slows down all future changes, exactly because the person who wrote it didn't take the time to think about the future or even good architecture in the present. Every time I touch their code I have to almost redo their work because:

- It's impossible to understand what is going on (For example when people load huge json or sql files into a single test and you don't know what scenarios it covers because they are too lazy to write dedicated minimal test cases. Another common offense is a not well-defined data model that gets mutated everywhere in dynamic languages like Python or JavaScript where you never know what shape an object has or where it comes from)

- Their code allows so little extensibility that following the same "pragmatism" as they did and just hacking it in would exponentially increase the complexity

- There are no abstraction boundaries which almost surely means insufficient test coverage (which I have to amend before I can even start with my own work so that I don't introduce regressions) or the tests are coupled so closely to implementation details that any change necessary for my own work will require me to rewrite all the existing tests

You are fast because you pawn off your design work to whoever comes after you. Of course you have fewer defects: your colleagues have to put in more time to maintain the same level of correctness or else risk introducing bugs. That's the cost of bad maintainability. It's well known that changing existing code is harder than writing new code, double so if it is not written with care. Someone has to be the janitor to keep the cruft from accumulating and if it's not you, then everyone else has to pick up your slack. Ideally, every person in the team would continually clean up and always leave a code base better than they found it.

Lastly, I want to say that there is very important difference between simple and easy and simple doesn't come for free.

It sounds like you have personal experience in this. I hear you. From what I have observed through code reviews that I, or others, have looked at that doesn't happen with my code unless there is a very fundamental shift. Even then I try to keep my inputs and outputs clean so that pieces can be removed if need be. Of course I am not always as successful as I would like to be.

Now my first 2-4 years, that happened a lot, and I am not going to claim that I was fast. I was slow and either the code review would point something out that I would literary have to start from scratch or the next person that touched it would have to. This was actually the driver for me to focus more on simplicity. I basically discovered a direct correlation between complexity and how bad my code was.

To this day, if I catch myself doing something overtly complex it's a sign that my solution to the current problem is inadequate.

Ironically, my response would be the same if my code was trash. So I won't be able to convince you regardless.

I am confused why you are just taking the least charitable assumption based off OP's short post, and writing these long screeds just taking it as given that your assumptions are true.

I think they're just using this as an opportunity to rant about a personal situation.

My personal assumption is that anyone who thinks of themselves as a “10x developer” using the criteria he is using, probably isn’t. He only has 5 years of industry experience and I’m assuming at only one or two companies.

> I became a Jr Dev at 33 years old...

> I'm now 42 years old...

That's 9-10 years.

Do you know OP or have you seen their code? If not, you're making a lot of assumptions about them based on what they've said that are about as uncharitable as possible, to the point where it kind of seems like you have an axe to grind against OP.

I'm speculating here but I think that you have some major hurt and pain in your past and something OP said is triggering the memory(ies). There definitely are (many) people in the dev world that do what you're accusing OP of doing, so I'm guessing you've been heavily burned by that (as have I). I actually moved industries to get away from that. In my case it was the Java (especially "enterprise" java) culture/ecosystem/etc, and things got a lot better for me when I wrote off my deep investment in Java and pivoted to C++ and Ruby. I would definitely advise to move around early and often, both company and tech stack. Try out startup life if you've been in big corp, or vice versa. IME startup life is a lot more prone to the "ship it fast" pressure that results in terribly unmaintainable code, so a big corp where you can work on a core product (core enough that it gets plenty of investment and appreciation but not enough that it's the number one money maker, otherwise the Eye of Sauron will be omnipresent and encouraging bad practices). Life is way too short to be miserable at work, especially in a field like ours where hold a ton of the cards.

Seen the same thing happen with devs that proud themself to write “simple” code. Sure, every individual script they write is simple in itself, but in the end you have thousands of them that don’t compose or reuse at all, with lots of implicit dependencies and assumptions between them. Leaving the rest of the team slowed down, trying to make sense of it, careful not to add even more to the existing mess, while mr “pragmatic” keeps adding even more copy-paste. Basically the total complexity becomes bigger than the sum of its parts.

What you bring up is a common and concerning scenario. It would be better received without the accusatory tone, until you know more about actual situation. There are many good and fast developers that don’t apply to this concern.

I have worked with 10x developers. They exist. Not only are they more productive than most developers, their code is also more bug free than most developers. Get over it.

Reply on news.ycombinator.com