Like 0.25px

What is the point of that ?

You cant split a pixel, can you ?

AFAIK, 0.25px line is still printable.

Pixels are discrete, indivisible elements in the physical world (but not always the same size or shape), but pixels as units of measure are virtual math constructs and fungible.

I was jut thinking:

if i print a 10x10mm square at 300dpi (~120dpmm), thats 118px, right ?!

Thats roughly one paint dot per px. So maybe 0,5px would be printable at 600dpi ?

Not sure if i got the math right.

0,25 px=0,2 pt @90 dpi; you need to multiply the px with 0,8 to get to the point...

There are topics on cutting machines needing pdf-s with such thin lines.

Printing px values depending on the resolution -0,25 px can mean 0,635 mm @ 10 dpi.

Theoretically you can use viewboxes and transformation matrices -on groups- so then the 0,25 px stroke width would render wider.
(Couldn't reproduce the latter, maybe it was a "known bug" that got "fixed"?)

Espermaschine wrote:You cant split a pixel, can you ?

So it's about rendering a 0,25 px part of an image.
That depends on the type of anti-aliasing used.

Espermaschine wrote:Thats roughly one paint dot per px. So maybe 0,5px would be printable at 600dpi ?

The less the resolution the larger the "px" appears on paper.

But inkblots don't scale with dpi, my guess is that printers have a native resolution you cannot effectively set -unlike on old crt monitors.

So 0,25 inkblot wont get printed as per se.
But printers need to provide a larger resolution to provide the same appearance as on screen, due to inkblots don't mix.
Meaning 0,25 screen px may get printed anyway.

Lazur URH wrote:

Espermaschine wrote:You cant split a pixel, can you ?

So it's about rendering a 0,25 px part of an image.
That depends on the type of anti-aliasing used.

Right, you need a full pixel to make use of semi transparent pixels to display anti-aliasing. Didnt think of that.