Why Fractional coordinates is bad?

I’m using grid spacing:1, subdivisions:2 but I read fractional coordinates is bad. Why having a node on position x.5 is bad?

Which source says it’s bad?

It’s incompatible with Quark XPress’ PDF generation.

:frowning: but I read what upm different than 1000 causes bugs on Windows.

I didn’t hear about that?

Ok sorry, the bug was in Mac OS 10.10 and seems to have been fixed in 10.10.2

So you’d advice to just convert from 1000 upm to 2000 upm and get back the grid to spacing:1 and subdivisions:1?

Unless you are doing a very finely detailed font, you really don’t need anything more than a 1000 UPM set to 1/1. For a very finely detailed font a UPM of 2000 should suffice.

I find 1000 ×2 upm makes anchors position better suited for smooth curves, especially when interpolating quite extreme masters. When I worked with 1000 ump with subdivision:1 before, I found it not precise enough for the font I’m working on now.
Is there any rendering/compatibility issue or caution while working on 2000 or 4000 upm? Does it work on Mac, Win, webfonts, etc?

Problems start around 3000. This is mentioned in the handbook.

Ok thanks!

Sorry, so if I understand well, because of Xpress, it is also bad to use
"Grid Spacing" = 0; as parameter into instance interpolations. I found it useful as it can get smoother diagonals…

That’s right. Or you attach a label that says incompatible with XPress.

what a pain :frowning: isn’t planned from Quark to fix this bug in Xpress pdf generator, or it “is a feature”?

If enough people complain at Quark, they may fix it.