Native FBInk: Another small tool to print crap on your screen!

[PoP]

In line

Originally Posted by NiLuJe

@PoP:
...

I wouldn't mind a full set of lowercase letters, when you have the chance .

Serious? I tried once and it didn't look good. I wanted to stay faithful to the 6 atoms in a 4x4 grid. Amazing that it could render legible upper case. Perhaps if I introduced the "box characters" as new atoms for the lower case

Would using the same A-Z glyphs for a-z satisfy you?

(Some glyphs might need "kerning" tweaks, at a quick glance, some digits, for instance: printing "1234" makes it somewhat obvious ^^).

Not sure I understand. Again, Isn't it keeping up with the blocky style?

EDIT: Also, the last four characters have the wrong codepoint .

Corrected result of a manual edit.

Found a minor bug in the conversion script while working on this, and this lays the groundwork for making BDF conversions even easier now . Everybody wins! .

EDIT²: Yuuup, BDF conversions are a definite go! .

Nice, I will use that format then.

BTW Want to downsize to -F MICRO? The minimum BDF font size I found was 6x4 pixels.

EDIT: And titty went as low as 5x3!

[NiLuJe]

* Yeah, a scaled down version of the same glyphs would do just fine for the lowercase stuff .

* See the attached snapshot . I get the idea, but (some of) the digits are a bit all over the place (mostly 1, 2 & 3).

I'm considering the lack of padding a feature and not a bug ^^. Might be worth experimenting with a tiny bit of padding for the lowercase stuff, though? But I don't really have a strong opinion on that one .

* Yeah, there's a lot of really tiny BDF/PCF fonts (f.g., creep is 4xsomething, IIRC). Unfortunately, I can only use stuff with a *width* that matches the width of a C data type in bits (i.e., 8/16/32/64). If there's one column of difference one way or the other, than can usually be fudged w/ gbdfed, but even then, turns out that severely limits our possibilities .

(Plus, in our case, 8x8 is already mind-boggingly tiny on a >250dpi screen ;p)

[PoP]

* The scaled down version of the A-Z glyphs could (for the lowercase stuff) consist in A-Z (still upper case) but rendered with the box drawing characters. That way, it somewhat preserves the the blocky style look while still distinguishing a-z. in your screenshot, I liked the thinner A-Z as a start for a-z, à la small caps.

* I think shifting the 1 one column right, and the 3 one column left would fix the scattered digits.

* In my opinion, for that font, the lack of padding helps the style.

* "data type in bits 8/16/32/64", Got it. (Plus, in your case, no need to go below 8x8 if only to create a special effect by scaling up with -S $PARAM)

[NiLuJe]

* Sounds perfectly fine to me, go crazy! .

* Yeah, something like that .

* Agreed! .

* Turns out I was a bit hasty in that first assessment ^^. The only actual hard rule is <= 64 .

EDIT: I whipped up a BDF version in this commit (via Unifont's hex2bdf script).
Might make it easier to work on, if you have access to gbdfed .

[PoP]

You were right. I am going crazy, and nowhere. This could be long.

Thin glyphs where lines touch borders don't mix well with the blocky ones.

Might as well extract a-z from any other font if the thin lines won't reach borders. Then the style suffers.

Hollow triangles made with 1 pixel diagonals and straights instead of filled ones look ghostly.

Scaled down 4x4 filled triangles in a 8x8 grid reduces their weight OK but they obviously won't touch anymore where they should at junction points.

Box-drawing ...well, was meant for boxes.

In the mean time, my unsatisfying suggestions:

#1: reuse the A-Z glyphs for a-z always displaying text in capitals, BLOCK style obliged.

#2: if absolutely need to distinguish lower from capitals, invert A-Z pixels to display a-z.

...Sorry PoP is, so far, a failed font designer.

[NiLuJe]

Yeah, I was thinking about #1 if all else fails, too... .

[NiLuJe]

With those few aforementioned tweaks in .

EDIT: And with a tweaked question mark, so it looks less like a stretchy seven...

[PoP]

Relieved. You're the best!

[NiLuJe]

Released v1.3.0 .

This one comes with a rather large bunch of changes .
I won't repeat myself, see the Kobo thread.

On the Kindle front, the builds are now unified, a Kindle build now handles every device properly .

[PoP]

@NiLuJe Thanks again.

Spoiler:

fbink -cmM -F block -S 1 BLAM
fbink -cmM -F block -S 2 BLAM
fbink -cmM -F block -S 3 BLAM
fbink -cmM -F block -S 4 BLAM
fbink -cmM -F block -S 4 BLAM!

[NiLuJe]

And released v1.4.0, a bit early, because I'd stupidly broke the Kobo package in 1.3.0 ^^.

This one has some more text rendering performance improvements, as well as the ability to do pixel-perfect positioning of text, much like you could for images.

[PoP]

Originally Posted by PoP

You were right. I am going crazy, and nowhere.

Eureka.

I had an idea for the lower case characters. Reduce the atomic triangles from 8x8 down to 4x4 and center 16x16 glyphs in a 32x32 grid.

Voilà: ꜱᴍᴀʟʟᴄᴀᴘꜱ in a blocky shape identical to the upper case.

@NiLuJe Do you think this might work? Attached are the .# drawings of a to z, plus a correction to the middle leg of the M:

M+atoz.txt

[NiLuJe]

@PoP: I'll try it out .

I'm not seeing any change to that M, though?

EDIT: Here's what it looks like in gbdfed (compared to the W, which doesn't look off either).

EDIT²: And here's how that looks.

(NOTE: I had to plug some holes ).

EDIT³: w was missing a slope .

EDIT⁴: o was missing a cap .

[PoP]

Inline

Originally Posted by NiLuJe

@PoP: I'll try it out .

I'm not seeing any change to that M, though?...

OK then, perhaps problematic only in my internal draw files .

EDIT²: And here's how that looks.

Ahhh. Satisfaction.

EDIT³: w was missing a slope .

ACK.

EDIT⁴: o was missing a cap .

ACK.

[NiLuJe]

@PoP: I may have noticed something wrong with the M earlier and promptly forgot about having tweaked it, but that really doesn't ring a bell (and I can't find anything in the repo's history either) .