[kwlug-disc] UNIX in 1982

[Ronald Barnes]

Khalid Baheyeldin wrote on 2024-10-15 10:56:

> In previous threads, we discussed how newer
> generations of developers and techies invent
> new ways of doing things, while such ways
> contradict wisdom that has very good reasons
> of why it is there:

When things change, sometimes those "very good reasons" aren't as 
compelling as they once were.

And tech changes rather often. It's not 1982 anymore.


> Examples include :
> 
> - shared libraries impact on resources, maintenance,
> and security, vs. distributing statically linked mega
> blobs

Mini "blobs" are okay though? That seems like a loaded term, "blob".



And should your concerns on resource usage take precedence over 
developers' time and effort required to optimize their (*free*) software 
to your constrained environment?

When it's reasonably inexpensive to acquire necessary resources (RAM and 
disk space is quite cheap vis-a-vis developers' time).



If developers are giving away software for free, it's their choice on 
how to compile it.



> - giant programs that ignore the "do one thing well"
>   principle,

Such as?


> - structured files vs. simple string of bytes

Which is the desired scenario here?

I'm thinking JSON / YAML are new, therefore bad?



> - The tendency to rewrite every old thing because
> "old is bad" and new is better ...

Or the tendency to cling to every old thing because "new is bad"  and 
"old is better".