On Mon, Apr 7, 2014 at 5:46 PM, Brian Wolff <bawo...@gmail.com> wrote:
> No you don't get more consistency by moving back to an experience > where you let the browser determine fonts. However you do get a > situation where things are more likely to work for non-latin scripts > (and other issues that have been brought up, if I understand > correctly). Consistency and actually working for all scripts are > separate goals. If you can't satisfy both, you're going to have to > make one take higher priority than the other. I personally consider > the "Actually working for all languages" to be much more important > consideration than consistency. > I totally agree. I don't see how there is any indication this is functionally broken or a major regression across languages, keeping in mind the necessity of ULS et al still. What major language-related bugs have been raised that would not be present regardless for most default sans-serifs? I see some cases, such as CJK, where users may not prefer the serif headings, since serifs look closer to a brush script in those languages than they do in Latin languages. That doesn't seem to be what we're discussing though? When it comes to the body font settings and language support, we've been able to remove some local overrides. Some others, like Persian, had Common.css overrides long before the new version was released. The general state of affairs before *and* after is that there aren't great-looking, freely-licensed fonts with support across all languages and which are widely installed on our most popular OS/device combos. We're making incremental improvements here. _______________________________________________ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
