2 ms·
> I always find this such a weird claim. Sure, lots has happened in the past 5 years, but there's still tons It depends a lot on what you are working on. If y
by ThinkBeat 4y ago
> I always find this such a weird claim. Sure, lots has happened in the past 5 years, but there's still tons
It depends a lot on what you are working on.
If you work on a big legacy system, you can probably walk in after 5 years and
pick it up.
Same C on an established embedded platform.
And certainly, Cobol on a mainframe.
Perl you are good as well
It seems like if you writing front - end code shit is changing on a monthly basis , framework come, disappear, new hotness, cargo culting.
Or at least I think that is what a lot of the cursing I have heard from people who do it.
Myself i have never down that road but it seems chaotic.
I have worked a lot with C#.
It has changed so much over 5 or 7 years that if you are all up
on the latest changes you can write code that is not easy to read
for someone a while back. (Here I am talking language changes)
Of course the .Net framework has changed a lot, .core now it.
You can pick it up but it would take work.
Really it depends on what, where and how.
- mbrameld 4y ago> It seems like if you writing front - end code shit is changing on a monthly basis , framework come, disappear, new hotness, cargo culting. We're taking about from 2018 until now. React was already 5 years old back then. If one was a productive frontend dev in any of the popular stacks of 2018, what exactly do you think they would be missing in order to be productive today?