Hi!

It's very nice!

3x4 (please a tag), with a square look (between semi-bold and bold), balanced type color (few 'inkbloats'), high and low bars mixed, very decent lowercase (3 x-height with 3 baseline hacks), a complete set of symbols/punstuations (I'll see later if it fits US-ASCII, and try to remember your own/cool solutions at this scale, perhaps in a future comparison with mine). One (unnecessary?) 3x5 glyph is outside your matrix (for extra line spacing?). There are many serifs that match the monospace style (max width of glyph, i.e. 3-dot wide), but you included a couple of proportional glyphs (from a variable width as a base for your design? it is possible to 'fix' both). The uppercase N is different from what we often get in 3x4 (maybe it'll be a very known alternate, one day). The $ looks like %. # is a kick into touch. ~ is pretty small (thus cool). % ~ i! (mono abuse). Several punctuations/symbols would not be viable without monospacing either. Result: a font born to reside on a computer. Designs for the real world are quite imaginable (mine included). I feel you had fun, and are skilled. WELL DONE!

Oh, so:

• 'tag': Fave Tags field (see under Monospaced in this page);

• 'type color' is the gray (in French): with even distribution or not (e.g. _XM where you can get the 3-color gradient);

• 'x-height', 'lowercase' yield no difficulty(?);

• 'baseline hack' (personal jargon) is e.g. a shifted/raised baseline (location: y-1, that is your lowercase trio here: gpq);

• 'matrix' ~ grid (orthogonal cells in your case of simili-LEDs);

• 'monospace abuse' (personal jargon): see there for an example in 4x3;

• 'kick in touch' (personal metaphor): any filled grid, single (like your #) or repeated (as in Microsoft's "Small Fonts" typeface), that is not retrievable from the context (i.e. the block size or/and location could not help);

• 'real-world fonts' (personal concept): fonts (most likely proportional in width and very low-resolution) that e.g. unequipped humans (even many animals?) can use (read, write, manipulate, shape/build etc.) with a minimal memory/brain and effort/energy/resource (it's all about ecology, Mr T.!).

Please tell me if there's any other puzzling term in my comment (for the specialists, OK).

---

I paste you the metadata of the illustrated PNG above (uploaded in the next message), bye!

Dpla's suggestions for dollar and more signs, to "12Squares", a 3x4 dot-matrix monospaced and +/- US-ASCII pixel font released on 2017-09-14 by Lou D. (Lou FoxyBun) on FontStruct.com.

3x4 text characters: $ (row 1), # (row 2), mix (row 3):

* Pertaining to the first row, this is an incomplete set of the noise distribution techniques (a.k.a. custom binary order) in pixel art devised by dpla since 2013 (especially from his list of 3x3 alternate glyphs for the documentation of his very low-resolution fonts project);

* Pertaining to the second row, the extra colors highlight dpla's main suggestions or comments in the context of Lou D.'s already fontstructed choices;

* Pertaining to the third row, dpla added a few flashable glyphs, just for fun.

https://fontstruct.com/fontstructions/show/1436227/12squares

dpla.fr/fonts

2017-10-20 dpla

'Monospace abuse': without this 'hack', "%" would not be viable (it would duplicate "i!" in proportional). Now that you got it, your following difficulty relates no more than to the common tough choices we all have to make in very low-res {-if my English still makes sense here-}, i.e. the '15 puzzle' way of designing by time-consuming trials and errors (in other words: 1 glyph mapping = 1 code point).

I needn't go further in my (already complete) suggestions, since it's more or less always a matter of:

1. STYLES (personal and basic choices: several representations of our Latin characters coexist, but all are not easily/evenly shrinkable amongst the small grids - the font weight or the related pseudo-anti-alisasing must not be ignored, even at this size);

2. PRIORITIES (rules are to be applied -quite preferably- by subsets: see what I mean from my latest AGH project (completed), seen as an introduction to this field before digging deeper into this promising world of 'micro' text characters);

3. LOGIC (the more repetition, the more coherent: except for very fanciful -3x4- glyphs that are to be avoided, you shorten the learning curve from this precaution - b.t.w., your current "$" could not be kept if you took care of this point, since a small lowercase "s" plus a "\" symbol across it are two effects-by-logic that don't repeat, which make this glyph unique in its subset, thus an aberration, not mentioning the fact that your (runner-looking) "$" creates an ambiguity with your "%" and "#", and not the contrary).

So,
a. you are planning a new "$" [before | after 1 or 2]:
▓░░ | ███   ███
░▓▓ | █░░   ██░
▓▓░ | ░░█   ░░█
░░▓ | ███   ███

b. where your current "52SZ" look like:
███   ██░   ░██   ███
██░   ░░█   █░░   ░██
░░█   ██░   ░██   █░░
██░   ███   ██░   ███

c. and you are thinking of a new "S" [before | after]:
░▓▓ | ██░
▓░░ | ░█░
░▓▓ | ░█░
▓▓░ | ░██

Then, //to me//:
a. both new choices look better (glyphs #21 or #24 in my image);
b. but your current "S" looks like "Ç" (I'd see e.g. #30 instead);
c. this glyph looks like a standard (though sometimes disconsidered, it adds to the repetition, i.e. ± logical mirrors, thus it's ease the readability).

Again, it's not much _my_ opinion that counts in _your_ design, eventually.

Whenever you make durable choices, solid hopefully (e.g. like 2-dot-wide numerics - not featured in "12Squares"), you'll have to struggle to make them outlive the other priorities. I leave you with this question: what are them, in you font, precisely? (all the subsets should be ordered by descending choice of design at this minimal size, else we would not discuss about the ambiguities that arise inevitably between them, e.g. your new "$" version 2 that would collide with your current "5", i.e. the difficult-to-follow ordering: symbols > numerics.)

“Other people are better placed to judge the readability of it”: I repeated this in 2013, pertaining to 'micro' glyphs and the final goal (not a personal/experimental/autistic mapping). Now in 2017, all I see, is always more incomplete fonts (not even US-ASCII, often just words to look retro, commercially), and 5-dot of width ("m" and "w", which is too easy)… so, 99 % not very well thought-out. (The IQ might lower according to what is said in the French news, but I have noticed the lack of concentration, esp. on this recreative site. A font that is designed in one day or week, cannot compete with the ones that use decades of search/comparisons and trials and errors, IMO.)

---

Another tip: it's only once you are fine with your choices that you can tweak a few glyphs, in general according to what is already chosen by the community. Yes, this kind of design can get frustrating, but I know you master this field of patience!