Improving KDE Plasma Performance on Low-Resource Systems

[Reindl Harald]

Am 21.08.26 um 20:18 schrieb Aaron Rainbolt:
> It's worth noting that modern web browsers and the modern web as a
> whole is very resource-intensive. One tab with a modern chat
> application (Element) can consume a gigabyte of RAM on its own.

which shows the bullshit in modern IT at it's own

with 1 GB o RAM i had running an Win2000 Server with Corel draw, 
photoshop, VMware with a small Linux system playing music in the 
background while try to learn software development with "Miscorosft 
Visual Basic 6"

don't get me wrong but modern software *by far+ don't have that much 
functionality and comfort compared with the hardware ressources we a 
throwing at it

> If you're trying to do serious work, the reality is you may simply
> need a faster computer. You could probably use the device you're
> looking at with a lightweight desktop like IceWM, and it can still do
> things like chat on IRC, compose text documents, and run a text-only
> web browser or perhaps a very lightweight, incomplete browser like
> Dillo or Netsurf, but what you're trying to do is simply not something
> your system can handle.
that's nonsense

after 25 years i still need to see any piece of complex software which 
can't be optimized by MAGNITUDES

i have written enough stuff in the past to know that a few minutes of 
thinking and chnage a few lines of code can make things 2, 5 oder 10 
times faster or consume only half or the memory with the same outcome

it's always a tradeoff between throwing money on hardware or human 
ressoucres on a problem but in case of re-used code even small 
optimizations can have a huge impact on a whole ecosystem

a single line of code within a loop where the result before would be the 
same can make things *100 times faster* and that single line outside the 
loop not only itslef is 100 times faster - inside this call you can have 
dozens of functions with the same impact becuase of loops in itself

often you can#t measure such things, write sofwtgare which works 
perfectly and fast in all test cases until you don't have 100 but 
100.000 iterations and then a few lines of code changes can make the 
difference of taking 1 second or 10 minutes, faced it often enough in my 
daily job and there are cases where you can gain *1.000 times of 
performance*

imagine what that means worldwide in case of energy alone multiplied 
with millios of machines - it won't save the world but i don't see a 
reason to buy new hardware and consume tons of energy without a good reason