- From: John Hudson via GitHub <sysbot+gh@w3.org>
- Date: Sun, 16 May 2021 04:59:51 +0000
- To: public-css-archive@w3.org
> I also wonder if it's time for another axis/axes, like "ptsz" and or "pxsz", which can be more carefully implemented, and opsz can be deprecated. We’ve been down this road before with every optical size selection mechanism introduced into OpenType: the size GPOS feature, the OS/2 size range values, and now opsz axis. We’ve been down similar roads with vertical metrics and several other aspects of fonts, adding bits and flags and new structures to work around failures to implement correctly. And then everyone complains about how messy the OT spec is. > I also hesitate to suggest that, despite the work to reify the opsz definition as real physical pt, at some point the OpenType spec has to concede that implementations did something else and document what is actually implemented. See earlier discussion about impossibility of designing size-specific glyphs for unspecific sizes. So long as ‘what is actually implemented’ is 1:1 mapping of opsz values to non-physical units—instead of calculating between those units and the physical units of the type design space—it cannot be designed for. That is the issue for type designers and font makers. Now maybe it is simply going to be the case that CSS opsz is going to be a kind of endlessly adjustable fuzzy implementation, in which there isn’t actually any expected level of accuracy in design size selection, just something that might be somewhere in the ballpark, and type designers will stick to making size-specific designs primarily for print. And I am, perhaps surprisingly, sort of okay with that, because the unreliability of scaling anything on the web to be the same size in two different places seems a characteristic of the medium, so size-_specific_ design was always going to be a problem. -- GitHub Notification of comment by tiroj Please view or discuss this issue at https://github.com/w3c/csswg-drafts/issues/4430#issuecomment-841766776 using your GitHub account -- Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config
Received on Sunday, 16 May 2021 04:59:53 UTC