On 08/04/14 00:02, Steven Walling wrote:
On Mon, Apr 7, 2014 at 4:08 PM, Erwin Dokter erwin@darcoury.nl wrote:
I feel that I am not being taken seriously. Three times now I have indicated what is wrong with this solution, namely that a single font stack cannot possibly serve a global website.
I want to ask Steven and Jon how they plan on serving *all* the scripts and languages in the world in a *single* font stack. There is not a single font in existence that can possibly support all languages.
I think we actually answered this up front at https://www.mediawiki.org/wiki/Typography_refresh#Is_there_a_perfect_font_th...
Ultimately we're shooting for and getting a lot more consistency and control over the user experience here, for most users. That doesn't meant that it's perfect. There is definitely not a single font that is available everywhere that supports all languages. That's why it's a font stack with fallbacks. We definitely don't gain more consistency across the experience by moving back to a situation where the styles basically just define no style.
Again, this excercise is completely Latin-centered. Projects using different script have no choice but to override to their native fonts, and only Europe/Americas is left to 'enjoy' the new font stack.
To add on to what Jon said: we're going to figure this out in discussion with the communities. I don't think it's the case at all that users "have no choice" if they want readable text in a non-Latin script. To use CJK as an example: I actually was able to remove some local hacks that were necessary before the new version. We'll keep working on it. _______________________________________________ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Why? All this effort, and for what? You want consistency in what's given to the users, but users are not consistent. Their hardware is not consistent, nor are their operating systems, languages, eyes, brains, habits, or preferences. As was, the default sans-serif already addressed these inconsistencies, giving them something that worked for them. Why go to all this trouble when the problem was already solved? What's the point?
-I