Thanks! To me it still doesn't look like an adequate replacement for Times New Roman, Palatino, Bembo, Computer Modern Roman, etc.; it's still a monoline font with no serifs but the occasional slab serif.
I know de gustibus non est disputandum, but if you picked up a book in a bookstore and it happened to be printed in Intel One Mono, would you really think, "Oh, what a visual pleasure!" or would you think, "Oh, how unfortunate that 50 years ago they couldn't afford to properly typeset this book and so just xeroxed these typewritten pages!"? I would think the latter.
I've enjoyed the occasional old typewritten paper I've come across so maybe that is part of what I like about it :). I much prefer that to many of the issues I see all over the place on the web (like too small line height). I think there is a density vs readability tradeoff and Intel One Mono does well at lower density readable text (at least on screen). Although I think some proportionality can be done without loss of readability and hard to say what I would prefer in print.
I've been meaning to try Intel One as my main sans-serif font in the web browser and just changed it (adding it to the top of sans-serif in fontconfig). Looking around a bit I would say the one less than ideal aspect is the word spacing and setting that to -.3ch via stylus to test (unfortunately not a long term solution since it messes up other fonts) makes it look even better (particularly with justified text which gets quite unreasonable with the mono space as the starting point). Maybe I'll try to create a derived "almost mono sans" that just adjusts the space width.
I had a new idea as I woke up today about the layout problem of reconciling a character-cell grid with proportional text, but I don't have it fully understood yet. Basically the idea is that, if you know that a given span of text contains no internal horizontal alignment constraints—for example, a line of a paragraph, or a table-header column title—you can lay it out with internal proportional spacing.
If you run xterm with a proportional font, for example as
xterm -fn '-*-helvetica-*-r-*-*-18-*-*-*-*-*-*-*'
you can get a sort of crude approximation of this behavior; each string it receives from the process in the pty as a single operation gets drawn as a single string and therefore has reasonable horizontal spacing internally, but characters drawn one at a time get placed one by one at their properly horizontally-aligned location. So if you type "ls -al", the six characters show up far apart because they appear one by one; if you then type ^U^Y to erase and redraw them, they appear together as a single string.
(I kind of hate Helvetica, but because it's so proportional, it's the font that demonstrates this xterm behavior most clearly.)
xterm is doing this by accident, and so it does things like fail to erase characters that should be erased, use a far-too-wide width for each character column, and not tab to a requested column even when processing actual tab characters (like in ls -Ca output). But you could, I think, improve the behavior substantially with a few small changes.
First, you'd need some kind of explicit indication of which spans contained no internal alignment restrictions, instead of just using the stochastic clue of which characters showed up in a single read() call, which I think is what it's doing.
Second, you could allow each column of the character-cell grid to contract to contain only the characters it needs to contain, rather than giving them fixed locations; using a narrower-than-one-en width for the space character (as your -.3ch kind of does) would probably help with that.
Third, you could probably stretch and squish the actual letterforms themselves—TeX doesn't do this (though it does adjust interletter spacing), and hot lead can't, but medieval scribes did it constantly, and such "microtypography" is coming back into vogue. The place I see it most conspicuously is in the Android Heliboard autocorrection buttons displaying candidate words, though in a fairly crude and rebarbatively exaggerated form.
Fourth, certain glyphs (such as box-drawing characters) definitely have to be stretched to fill out the entire requested space.
I'm not sure this is a good idea, but it might work acceptably.
Thanks for the thoughts and encouragement :). Microtypography is interesting and contracting letters to justify would definitely make a lot of sense when starting with a near monospace proportional font. It looks like a nice keyword used in a few other ways as well in some interesting typography articles :).
https://github.com/intel/intel-one-mono/releases