Hacker News Reader

Comment by dsr_ | Hacker News Reader

Does anyone else think there might be a significant security value in not allowing a web browser access to a font associated with messages from the operating system?

View on HN

Unless the font is particularly distinctive to the most casual reader, such an approach seems like pseudosecurity. San Francisco is not distinctive enough from other well-designed serif fonts that people will intuitively notice a spoofed message.

Besides, malware authors generally cannot be expected to follow copyright law. They can just steal the system font OTF and convert it to a web font, or render text as an image (which may not even be illegal), or any number of similar things.

Agreed on the first point. Your second doesn't make sense to me, since any situation where using San Francisco is going to give an extra sense of authority, is also a situation where it's already available to use.

Hmmm, I guess your argument is that Apple could prevent San Francisco from being usable as a font in even a desktop word processor, so you can't make an error message in that font and screenshot it?

It seems like this boils down to DRM on the font -- only making it accessible to Apple-signed applications (not even to third-party apps, since I can write a third-party app that pops up a custom error message and then take a screenshot of my app). This is untenable for the usual reasons that DRM is untenable.

If you are spoofing content, web fonts already have you covered. Assuming that people even notice the font enough to take it as a security cue.

It might be possible to do OS fingerprinting based on the font metrics, bypassing a spoofed UA.

This can already be done. This doesn't make things any "worse" than they already are in that regard.

Cf. http://www.guanotronic.com/~serge/papers/fc15-fonts.pdf & http://pet-portal.eu/files/articles/2011/fingerprinting/cros...

Reply on news.ycombinator.com