I tried to optimize Google web fonts query to include basic latin set + some of latin ext characters vital for my native language (Czech).
https://developers.google.com/webfonts/docs/getting_started?hl=cs#Quick_Start
The link above states that I can modify query to include only some characters to make it significantly lighter. So i tried those characters:
aábcčdďeéěfghchiíjklmnňoópqrřsštťuúůvwxyýzžAÁBCČDĎEÉĚFGHChIÍJKLMNŇOÓPQRŘSŠTŤUÚŮVWXYÝZŽ.,?!;/-_:"'|()[]ˇ+*@#$%^&¨®°©
And the query looks like this( because all the "unusual" characters have to be html escaped):
The final result looks like normal and it is only 23KB instead of standard 45KB (with full latin ext charset). The problem is that on some computers, some characters are not properly loaded - they are rendered in Arial( for example "Ě" in word "ODPOVĚDI"). Can anyone help me where could be the problem or how could I trace it next time I see it? Or is it just because this feature is in beta in google web fonts?