Noodleface
A Mod Real Quick
- 39,755
- 18,443
IT/Software career thread: Invert binary trees for dollars.
- Thread starter Traak
- Start date
Vinen
God is dead
- 2,801
- 502
Because we've discussed this before
Also because... anyone who has worked for Defense knows this.
Sheriff Cad
scientia potentia est
- 34,331
- 83,375
If they really measure your code like this couldn’t you write code to generate pointless case statements? Like let’s say you wanted an if (x <40).. generate a case statement that has every value 1 to 40 on a different line, then the otherwise handles the > 40. Hell you could even do 40-1000 or something and have it go to otherwise over 1000. You could literally write millions of lines of code a day.Because we've discussed this before
Also because... anyone who has worked for Defense knows this.
- 1
- 1
TJT
Mr. Poopybutthole
- 47,154
- 129,646
On Friday I discovered another idiocy bomb from that guy, possibly the other staff eng in the org too.
I work in data infrastructure and one of the components of most orgs these days is Kafka. Now, Kafka is not the only one but it was one of the first as far as I know. These messaging services are used by various things to run applications, but they can also drive the org's decision making. At least how they are used today. Kafka is an immutable log of events created by an application to do something. Most of the time.
In the data space this means streaming all of the messages out of where they originate from Kafka and to some other location to be consumed by data processes that are not the application producing them. I get a number of "duplication" alerts from these. I start looking into them and the junior engineer on the team tells me that he just deletes them. These are not duplicates but logical duplicates. As in they existed across multiple partitions and offsets within Kafka but they do contain the same message body.
The junior engineers were told to delete them by manually going into the system and deleting them. This whole situation was caused by omitting the partition/offset metadata. So for years now whenever this happens they just deleted them. I spent some of Friday afternoon sitting down with the three junior engineers explaining basic Kafka functionality to them and how you can investigate this with tools like Offset Explorer if they are interested in better understanding. I also told them directly that if they are doing anything manually and regularly then that is wrong whatever it is. That if I had ever told my seniors 14 years ago that I had a solution requiring hours of manual workarounds to function when anything unoptimal happened they would have called me an idiot, asked why I didn't consider a list of issues, and that I should go try again.
I can understand these guys with less than 3 years in the industry didn't know any better and were just listening to their leadership. I am going to ask the staff engineer next week why he didn't recognize blatant retardation here and did nothing about it. He's a shitter too so I think I know my answer.
I work in data infrastructure and one of the components of most orgs these days is Kafka. Now, Kafka is not the only one but it was one of the first as far as I know. These messaging services are used by various things to run applications, but they can also drive the org's decision making. At least how they are used today. Kafka is an immutable log of events created by an application to do something. Most of the time.
In the data space this means streaming all of the messages out of where they originate from Kafka and to some other location to be consumed by data processes that are not the application producing them. I get a number of "duplication" alerts from these. I start looking into them and the junior engineer on the team tells me that he just deletes them. These are not duplicates but logical duplicates. As in they existed across multiple partitions and offsets within Kafka but they do contain the same message body.
The junior engineers were told to delete them by manually going into the system and deleting them. This whole situation was caused by omitting the partition/offset metadata. So for years now whenever this happens they just deleted them. I spent some of Friday afternoon sitting down with the three junior engineers explaining basic Kafka functionality to them and how you can investigate this with tools like Offset Explorer if they are interested in better understanding. I also told them directly that if they are doing anything manually and regularly then that is wrong whatever it is. That if I had ever told my seniors 14 years ago that I had a solution requiring hours of manual workarounds to function when anything unoptimal happened they would have called me an idiot, asked why I didn't consider a list of issues, and that I should go try again.
I can understand these guys with less than 3 years in the industry didn't know any better and were just listening to their leadership. I am going to ask the staff engineer next week why he didn't recognize blatant retardation here and did nothing about it. He's a shitter too so I think I know my answer.
Last edited:
- 5
Noodleface
A Mod Real Quick
- 39,755
- 18,443
Defense has the absolute worst engineers I've ever met (and a few really good ones that have gone there to retire). No one audits that shit.
Just as an example... I once had to solve a problem with some messaging that was sending a bunch of bytes and we needed to check specific bits. I put in some code to mask the bits and shift them so the end result was just something we could store in a variable. CS101 shit.
An engineer saw my code (he was a senior) and could not understand what I was doing. I'm talking like... A single mask and shift. I had to sit down with this dude for 30 minutes and explain masking and shifting of bits.
Last edited:
TJT
Mr. Poopybutthole
- 47,154
- 129,646
Noodleface
A Mod Real Quick
- 39,755
- 18,443
TJT
Mr. Poopybutthole
- 47,154
- 129,646
It's a great company to work at. I very specifically chose to work here for the nice low employee count and startup vibes. They just had idiotic pajeets in a staff and senior role for the past 5 years while the org was developing. Overall this company has very few of them. I am going to figure that the staff made sure the senior engineer was hired because they were both, purely coincidentally, from West India.
Problem is that they cannot hide their retardation in the code which is all I am looking at. Which is how I got that one faggot to quit. The staff is next on the list because he is just as lazy and retarded.
Problem is that they cannot hide their retardation in the code which is all I am looking at. Which is how I got that one faggot to quit. The staff is next on the list because he is just as lazy and retarded.
- 2
Noodleface
A Mod Real Quick
- 39,755
- 18,443
I suppose but I feel like I'm too old to deal with that shit now lol. Different perspectives
Haus
I am Big Balls!
- 20,558
- 82,917
Any of these systems can be gamed, often programmatically. It's one of the critical flaws in modern "Manage by Metrics" mantras. In most industries now there is this seeming motion towards having managers who don't actually know what the people they're managing really accomplish. Instead they build dashboards and metrics. And they want a system where they can look at those dashboards and metrics and see "the measure of the man". It fundamentally tries to break Goodheart's law, which states that "When a measurement becomes a metric it loses it's efficacy as both a measurement AND a metric." It's usually pushed hard when companies shift from being technologically driven, to being "spreadsheet driven".
When my company started using a "AI Usage Dashboard" as a leaderboard to determine who was "embracing AI" people quickly realized they could set up agents to run useless AI tasks and be darlings of the leaderboard, and since management didn't know what "Good effective AI usage" really looked like there was no way to validate.
Noodleface
A Mod Real Quick
- 39,755
- 18,443
Same thing happened here. We had a guy who ran a daily script that would do a ton of shit with AI then delete it all. This was pre-credit economics. It showed him as one of the top users of AI in our large company and realistically he used it 0%.
- 2
- 1
