The Bulgarian form of Cyrillic in YouTube Sans incorrectly reproduces some glyphs

Stefan Peev

As you can see from the picture, the Bulgarian form of Cyrillic in YouTube Sans inconsistently reproduces glyphs such as /т, /и, /й. The reason is certainly in the OpenType features and probably not difficult to fix. It would be great if someone could draw attention to the issue and inform the font authors.

John Hudson

This does not look like a font feature issue, but rather due to inconsistent tagging of text in the page content. Some of the text strings in the UI are treated as Bulgarian in a way that triggers the font BGR locl forms, while other strings are not so are displaying the default international forms.

John Savard

John Hudson said:

This does not look like a font feature issue, but rather due to inconsistent tagging of text in the page content.

Despite not having your level of expertise, somehow I am absolutely sure that you're 100% correct here.

However, in my ignorance of these matters, I tend to be of the opinion that it is very difficult to avoid this kind of problem in the real world. Let's say that instead of dealing with the application YouTube Studio, we were looking at the YouTube page for a video.

Some people commenting on a video are doing so in Bulgarian,  from the keyboards of their computers, and other people are commenting in Russian, from the keyboards of their computers.

And how is the web page supposed to know?

If one has a fancy and expensive word processor, you can invoke the OpenType features of your fonts. If one is using WordPad, they might as well not exist.

So my tendency is to think that while all these fancy OpenType features are wonderful in theory, at bottom it's really a stupid idea. What good are features that can't be used when they're needed?

But that is, no doubt, mistaken thinking. The people who really need to use those features will take the trouble to use a proper word processor (and they don't even need to shell out the big bucks, LibreOffice being free). Fonts should not go around popping up dialog boxes of their own - as if they even could - to fill in for features missing in applications.

But before there was OpenType, of course, the workaround was to have a font with a Russian version of the typeface, and another font with the Bulgarian version of the typeface. I think people should still be considering whether to use that workaround, or whether going full OpenType would really work better. Or offer both choices, as long as lightweight apps without OpenType support are commonplace.

Stefan Peev

I can for surely say that the problem is in the calt feature. If you turn off the calt feature, all errors in the Bulgarian form of Cyrillic disappear.

John Hudson

That would indicate a non-standard implementation of langsys variants, and still doesn’t make much sense of why the correct forms are occuring in some text strings and not in others.