I am bewildered by the amount of character typography experts seem to be able to discern from (what to me are) these plain little serif-less monotone Latin characters that litter our screens. I do notice when things are very wrong, but I'm sort of puzzled by the amount of design and individuality that is expected to be incorporated in such a small space. How much variance is possible under these types of constraints?
In terms of body text, typography isn't meant to attract much attention. The common reader shouldn't notice subtle differences. Bad typography is easily noticed and feels wrong. Good typography communicates as it should and doesn't command attention. Readers might not consciously notice differences, but they might feel them. Some poorly-designed blogs instantly make me feel claustrophobic. If they opened up the line-height on the body text it would instantly feel and read better. In the same way smaller differences in a font design can change its feeling, character and legibility.
Type designers spend much time adjusting subtle nuances of a font to evoke a certain feeling as a whole.
Intuitively, I think similarly--and then I put two pieces of the same text next to each other, one in Helvetica and one in Roboto, and I immediately sense a subtle difference in personality that I can't put my finger on but I know is there.
Roboto has slightly looser strokes and shorter tails than Helvetica, but their letter shapes are very close. Here is a diagram showing Roboto overlaying Helvetica:
I'm surprised nobody has mentioned anything about CJK glyphs. I still use Droid Sans as my terminal font, both because I like to program in it and that it had really good CJK support. I haven't read anything mentioned about that with Roboto - I hope they didn't end up ignoring that.
I appreciate the thoroughness of the analysis, even as my eyeballs started to swim amidst the typographical lingo. The article doesn't so much 'explain' the font, though, as explain the constraints under which the font was designed.
Even in understanding the constraints, I can't help but feel that Roboto is a font that's outsmarted itself. All of this hyper-sensitivity to the finer points of smartphone typography is great, but I think it may be achieving effects so subtle as to just not be there for most of the audience. In the meantime, it punts on actually providing a detectable personality (at least IMO). Helvetica has a personality, and eighty years of baggage and connotation that goes with it. A frankenfont is never going to have that.
The subtle thing you shouldn't notice is readability (at multiple resolutions), I think it works well. Android can't use Helvetica anyway for licensing reasons, so they need an open source font; Roboto is Apache licensed.
In Roboto's defense (If there ever was one) it really is not fitted to be displayed on desktop OSes. Of course that kinda goes against Fleishman's argument that it is designed to be versatile across dpi configurations.
The article mentions "low 100s of ppi to 300 ppi," which is about the range for smartphone screens these days. Most desktops and notebooks still have displays at around 96 dpi.
A few paragraphs in to the article I was wondering what font the article was written in. I figured it might be Roboto and sure enough, a quick trip to Web Inspector shows it is.
I'm not a fan of Roboto personally as each character feels too narrow, squashed, and makes for tiring reading. I frequently check what fonts are in use when it's something unusual. The most beautiful by far that keeps cropping up is Proxima Nova.
All these rationalization still can't change Roboto as a whole is not well put together and has severe identity crisis. When your team look disjointed and at odds with each other, really there is no win in technicalities.
Besides, fonts are art with utilities, you can try to repaint Starry night again and again (each time more modern and technically advanced), Van Gogh is not perfect should not be the reason. I find Glenn Fleishman faulting Apple for not having an episode of Not Invented Here Syndrome rather baffling.
