Hacker News Reader

Comment by erulabs | Hacker News Reader

The "code that directs Social Security payments, tax returns and other payments owed to Americans" is "COBOL and ... it has been such a struggle to maintain it [because it was] developed for decades with very little documentation"

But having a SpaceX engineer start a rewrite is... bad?

I freak out when some minor component of my company becomes abandonware. When no one can or wants to read or work on the codebase, it's _time for it to die_. Either this good news, or you have no experience maintaining existing critical software infrastructure.

As for the "unelected" argument... Did anyone ever elect any of the software engineers who worked on this codebase? It has always been federal employees appointed by federal agencies. Some acronyms changed.

View on HN

He may be "SpaceX" (if that is supposed to mean super talented), but having a maximum of 1 weeks worth of knowledge about a system before starting to rewrite it seems extremely unsecure. This isn't a disposable rocket, this is a system that millions of people rely on.

Q: Is that what's happening here? That SpaceX engineers are rewriting the Social Security payment system in some modern programming language?

The article doesn't really give any good details. We don't know enough to say what the process is. We're just supposed to be upset that young bright engineers are working on the system that until a month ago everyone would have agreed is a pile of garbage.

I'd love to know more about what the plan is - but as a older software engineer "young smart kids attempt to start maintaining ancient abandoned but important codebase" is good news, not bad news. It's what we beg for in Slack all day long.

Do we know that guy is bright? And do we beg for that? I was 25 years old and thought I can do everything. Now I know better and would definitely not trust myself going into such a system. I was in many companies and especially if it comes to payment systems I see the same mistakes over and over again.

If there is atleast a good supervision, then sure it might work.

> Do we know that guy is bright?

One of the other guys won the Vesuvius Challenge, so I’d say probably.

'everyone would have agreed" - no, please provide any basis to that or you're veering into the "common sense" and "many people, some straight from central casting with tears in their eyes" BS

A month ago you would have argued that the computer systems backing our tax return filings were in a good shape? Fair enough, but color me shocked.

Point me to anything that shows they were bad Anyway, DOGE was supposed to be finding inefficiencies not rewriting the codebase, how much is that saving?

You know that asshole that joins a team and says "wow this code is bad, we should scrap it all and rewrite?" And if he gets his way, how he fucks off somewhere else before the rewrite is completed and leaves everything in a total mess?

You know that's consider a bad thing in serious software engineering teams, right?

As with all engineering, it depends. If the code is bad, and they do complete the task, then good engineering. If the code isn't bad, and they don't...

Armchair quarterbacking engineering decisions with little to no information is not something I'm in the habit of doing.

Reply on news.ycombinator.com