SkillsPython, R, Machine Learning, Scrapy
Joined devRant on 9/20/2019
Do all the things like ++ or -- rants, post your own rants, comment on others' rants and build your customized dev avatarSign Up
From the creators of devRant, Pipeless lets you power real-time personalized recommendations and activity feeds using a simple APILearn More
After over 20 years as a Software Engineer, Architect, and Manager, I want to pass along some unsolicited advice to junior developers either because I grew through it, or I've had to deal with developers who behaved poorly:
1) Your ego will hurt you FAR more than your junior coding skills. Nobody expects you to be the best early in your career, so don't act like you are.
2) Working independently is a must. It's okay to ask questions, but ask sparingly. Remember, mid and senior level guys need to focus just as much as you do, so before interrupting them, exhaust your resources (Google, Stack Overflow, books, etc..)
3) Working code != good code. You are an author. Write your code so that it can be read. Accept criticism that may seem trivial such as renaming a variable or method. If someone is suggesting it, it's because they didn't know what it did without further investigation.
4) Ask for peer reviews and LISTEN to the critique. Even after 20+ years, I send my code to more junior developers and often get good corrections sent back. (remember the ego thing from tip #1?) Even if they have no critiques for me, sometimes they will see a technique I used and learn from that. Peer reviews are win-win-win.
5) When in doubt, do NOT BS your way out. Refer to someone who knows, or offer to get back to them. Often times, persons other than engineers will take what you said as gospel. If that later turns out to be wrong, a bunch of people will have to get involved to clean up the expectations.
6) Slow down in order to speed up. Always start a task by thinking about the very high level use cases, then slowly work through your logic to achieve that. Rushing to complete, even for senior engineers, usually means less-than-ideal code that somebody will have to maintain.
7) Write documentation, always! Even if your company doesn't take documentation seriously, other engineers will remember how well documented your code is, and they will appreciate you for it/think of you next time that sweet job opens up.
8) Good code is important, but good impressions are better. I have code that is the most embarrassing crap ever still in production to this day. People don't think of me as "that shitty developer who wrote that ugly ass code that one time a decade ago," They think of me as "that developer who was fun to work with and busted his ass." Because of that, I've never been unemployed for more than a day. It's critical to have a good network and good references.
9) Don't shy away from the unknown. It's easy to hope somebody else picks up that task that you don't understand, but you wont learn it if they do. The daunting, unknown tasks are the most rewarding to complete (and trust me, other devs will notice.)
10) Learning is up to you. I can't tell you the number of engineers I passed on hiring because their answer to what they know about PHP7 was: "Nothing. I haven't learned it yet because my current company is still using PHP5." This is YOUR craft. It's not up to your employer to keep you relevant in the job market, it's up to YOU. You don't always need to be a pro at the latest and greatest, but at least read the changelog. Stay abreast of current technology, security threats, etc...
These are just a few quick tips from my experience. Others may chime in with theirs, and some may dispute mine. I wish you all fruitful careers!194
"You gave us bad code! We ran it and now production is DOWN! Join this bridgeline now and help us fix this!"
So, as the author of the code in question, I join the bridge... And what happens next, I will simply never forget.
First, a little backstory... Another team within our company needed some vendor client software installed and maintained across the enterprise. Multiple OSes (Linux, AIX, Solaris, HPUX, etc.), so packaging and consistent update methods were a a challenge. I wrote an entire set of utilities to install, update and generally maintain the software; intending all the time that this other team would eventually own the process and code. With this in mind, I wrote extensive documentation, and conducted a formal turnover / training season with the other team.
So, fast forward to when the other team now owns my code, has been trained on how to use it, including (perhaps most importantly) how to send out updates when the vendor released upgrades to the agent software.
Now, this other team had the responsibility of releasing their first update since I gave them the process. Very simple upgrade process, already fully automated. What could have gone so horribly wrong? Did something the vendor supplied break their client?
I asked for the log files from the upgrade process. They sent them, and they looked... wrong. Very, very wrong.
Did you run the code I gave you to do this update?
"Yes, your code is broken - fix it! Production is down! Rabble, rabble, rabble!"
So, I go into our code management tool and review the _actual_ script they ran. Sure enough, it is my code... But something is very wrong.
More than 2/3rds of my code... has been commented out. The code is "there"... but has been commented out so it is not being executed. WT-actual-F?!
I question this on the bridge line. Silence. I insist someone explain what is going on. Is this a joke? Is this some kind of work version of candid camera?
Finally someone breaks the silence and explains.
And this, my friends, is the part I will never forget.
"We wanted to look through your code before we ran the update. When we looked at it, there was some stuff we didn't understand, so we commented that stuff out."
You... you didn't... understand... my some of the code... so you... you didn't ask me about it... you didn't try to actually figure out what it did... you... commented it OUT?!
"Right, we figured it was better to only run the parts we understood... But now we ran it and everything is broken and you need to fix your code."
I cannot repeat the things I said next, even here on devRant. Let's just say that call did not go well.
So, lesson learned? If you don't know what some code does? Just comment that shit out. Then blame the original author when it doesn't work.
You just cannot make this kind of stuff up.87
My employer has a dev studio in Cali.
The office is gigantic.
It has amenities.
It has a stocked fridge full of iced coffee, energy drinks, and apparently wine.
All the devs have totally enviable hardware.
And they probably earn twice what I do, or at least 50% more.
Yet they write absolute shit, never test their code, and push broken updates every day, often marked as "ready for final testing." Their codebase is full of hacks and guesses and stale cruft and worst practices. I wrote a rant recently about one of their fuckups, which involved 18 million Facebook errors per. day. So that should give you some idea as to the quality of their code, and their level of can't-be-bothered.
Again, they make 50%-100% more than I do.
Their whiny lead dev is bloody lazy when it comes to building things correctly, and totally prefers to half-ass everything and complain instead. He probably makes 150% of what I do, doing like 25% as much work, and maybe 10% as well. Doesn't quite compare though, as he's a Unity dev, not a backend dev. So his work isn't as critical.
Bloody pisses me off.
"But their cost of living is higher!"
THEY SHOULDN'T EVEN BE EMPLOYED.23
Anyone use wireless earbuds (preferably not with a cord attached to the two)? Looking for an imformed decision. I usually prioritize sound quality over anything.9