Hacker News Reader

Comment by amelius | Hacker News Reader

It might be because in tech your knowledge of tools becomes worthless as soon as a new tools and ecosystems become fashionable, which happens quite a lot.

View on HN

I think that may be true if you focus on breadth rather than depth.

I work mostly on Windows, so maybe my case is special, but I'm still using knowledge about 8086 machine language I learned 40 years ago when I'm looking at a memory dump from a crash report today.

I occasionally dip into web-based stuff and although the top layers are different, underneath is usually HTTP / TCP / UDP / IP. What I learned with Ethereal in 1998 is still pretty useful today.

Truly new stuff doesn't come around that often. Maybe it's the difference between thinking about fashion and clothing.

I think the timeless skills you mentioned are much more useful for debugging, troubleshooting, and maybe overall architecture though. I couldn't for the live of me create a React app (because I stopped doing frontend long ago) but I can absolutely help with any of the networking, HTTP, or lower level stuff even if I'm just involved with the backend.

Yes, and if only 5% of work is architecture and advanced troubleshooting, while 95% of the work is plumbing in the language-du-jour then you see where the problem is.

The greybeard reaction to this is to pick a specialty, stack, and domain with less churn than the messier ones. Constantly treading water to stay afloat is exhausting and a waste.

Yup! Perl has kept me gainfully employed for 35 years and I'm still working - about to retire though (to spend some time dog and side-projects).

Perl is getting to be like COBOL - lots of lines out there still doing useful things but they aren't cranking out new Perl programmers like they used to. Lucrative if you can find the gig.

That said, I met a millennial COBOL programmer a couple years ago.

If you don't expect to be learning new stuff every day throughout your entire career, you are in the wrong line of business.

But it's not new stuff, mostly. It's just the same old stuff in a new package.

Yes, indeed that's true. But they get upset if you tell them that their genius new idea is just an old idea they didn't know about ;-)

That's precisely why it's not really that big a deal to learn the new stuff. If you know the fundamentals, learning the new tech is fine. It can be a little daunting at first but once you familiarise yourself with the new jargon and what value the practical differences bring you can apply your knowlege very productively.

No. Learning a new ecosystem is far more work and takes far more time than (so to speak) learning a few keywords of a new language.

Aren't you contradicting yourself?

Why? It takes time to master new ecosystems, even if what they offer is a complete duplicate of what already existed.

Not worthless... you can typically judge when a new tool has useful features, things to verify carefully before using or things that are likely a bad idea.

Tool advocates/salesmen may not like you though.

Reply on news.ycombinator.com