A couple of aesthetic adjustments about shantell-sans

[arrowtype]

I wanted to double-check this before moving ahead with changes, and I agree with my earlier gut instinct – the irregular O L N are just too tiny right now, even when the intermediates are being properly cycled through with calt. Depending on the letter order of a word, if one of these extreme letters appear in the middle or at the end of a word, they stand out too much. This, in turn, makes the full Irregular instances basically unusable. Scaling them back slightly (e.g. by using an interpolation between these and the normal styles) would make the whole thing more useful.

Using the Irregular axis, I come to two conclusions:

1. These small characters are best around what is currently IRRG = 65 (as of 537263e)
2. Another feeling of inconsistency comes from most of the glyphs getting bigger, and relatively few getting smaller. This is somewhat based on samples of Shantell’s handwriting, but I suspect the typographic effect might be improved by a slightly better distribution of inconsistency. I think I’ll start by making the T smaller, as that would have a fairly big impact (because T is so common in English), and still be pretty logically consistent with the system of small vs big glyphs. I’ll also make the A smaller, as this is another frequent character, plus it will help the style make sense in the phrases SHANTELL SANS and SHANTELL MARTIN.

A rough simulation of a smaller A and T:

[arrowtype]

I’ve made the interpolations with robofont Skateboard! 🎉 Super easy.

[arrowtype]

Oh no ... I’m making 3 alts, but the calt feature is only activating 2 of them. I will fix this.

[arrowtype]

I misremembered: alt3 does cycle in if a single character appears more than three times in a word.

Experiments so far:

• "Irrg A" – just a less extreme O N L
• "Irrg B" - less extreme O N L plus smaller A
• "Irrg C" - less extreme O N L plus smaller A I

Next experiment, "Irrg D":

• slightly less small A (it looks more extreme because it has more complexity/strokes than O)
• smaller D (more like O and C)
• slightly less extreme C

Final(?) experiment:

• slightly less extreme L

[arrowtype]

As a couple of records of the adjustments here, here’s an overview of the versions/experiments:

But here’s a better look, using some randomly-generated words. Previously, the irregular had a few extreme characters that really jumped out as being way smaller than the rest. Now – to my eye at least – things still look free/irregular, but without such standout tiny glyphs:

[arrowtype]

I’m going to move forward with the final experiment (or iteration?), version E. Here is the progression, if it’s useful later:

shantell-sans-irrg-experiments--2022_05_12.zip