Font Forum

There is a bug in the current version of FC.

The problem is exclusively with the 4 character sets that define the small letter i with diacritics.

When I pasted the glyph of the small letter i onto these four character cells and then complete the composite by clicking each character with the "Complete Composite Icon" on the tool bar, each corresponding small letter i glyph just disappeared while the diacritic composite appeared.

After some reboot and looking for obvious ghost manipulations, I discovered, by accident, that if I inserted the glyph in the dotless i character the glyphs in the four diacritic character cells for the letter small i appeared. The problem only occurred with the small letter i diacritic cells. Normally there are four of them by default. I had even deleted the default ones and re-inserted them but to no avail.

Subsequently I was able to complete the composites normally.

Obviously there must be some coding error that creates this depending behaviour on the dotless character i being filled with its corresponding glyph and this problem happened only with these four characters.
Al

Complete composites works by combining other characters to either complete the character correctly or give you something close which you can work with to make it correct manually.

If the characters it is supposed to use for the composites are not there it leaves a blank character, I don't think this is a bug. But Erwin would have to decide on this.

Since I did the coding, I can confirm that it is not a bug. If the composite was composed with accents + lowercase i, instead of accents + dotless ı it would be a bug. The result would look like this, which is incorrect.
Although it is not difficult to add a fall-back glyph of lowercase i for dotless ı, I don't think it would help users. What they need to do is add the dotless ı glyph to their font, not decompose every single composite with ì, í, ï, ĩ, î, ī, etc., and then delete the dot.

Similarly, if an accent is missing, they need to add it. The default New Font with outlines includes the dotless ı and some generic accents to make it easier for users, but they should redesign the accents and dotless ı to suit their font's design, and not just use one size and shape of accent for all fonts.

I appreciate your inputs but now I am more confused that before.

The way I complete composites is by copying and pasting each glyph into a composite cell from the main character glyph. I also do this one at a time. Then I click on the complete composite icon just to call the relevant diacritic. This is what I did for all the letters, both capitals and small letters. They all worked fine except the four small letter i in the way I described above. Throughout this process I had the dotless i character empty.

When I pasted the 4 letters i into the respective diacritic cells, one by one, and then called the complete composite for the relevant diacritic, the diacritic appeared but the pasted i disappeared. This implies that the four characters for the diacritic small letters i are dependent on the dotless i.

This is an anomalous dependance because it does not follow the same logic used to do the same process with all the other letters requiring diacritics.
Is this any clearer? In this contest therefore, it may not be a coding bug, but a logic reasoning and process bug.

algrass wrote: ↑Sat Feb 17, 2018 2:05 pmThe way I complete composites is by copying and pasting each glyph into a composite cell from the main character glyph. I also do this one at a time.

That step is not needed. Just use complete composites. You don't even need to use copy/paste from i to dotless ı — use complete composites, then make it simple and delete the dot.

algrass wrote: ↑Sat Feb 17, 2018 2:05 pm This is an anomalous dependence because it does not follow the same logic used to
do the same process with all the other letters requiring diacritics.

None of the other base glyphs: a, e, o, or u, have a dot, so they do not need special treatment. Examine any Windows system font and you will see that it has a dotless ı. Some languages (e.g. Turkish) also use the dotless ı as an alphabetic character on its own.

Bhikkhu Pesala wrote: ↑Sat Feb 17, 2018 3:38 pm Some languages (e.g. Turkish) also use the dotless ı as an alphabetic character on its own.

I've always regarded the presence of ı in a piece of text as a clear sign that the language in question is Turkish! I'd be interested to know if there are any other languages that use it.

Alfred wrote: ↑Tue Feb 20, 2018 6:43 pmI'd be interested to know if there are any other languages that use it.

Bookmark the site that I referenced. The Letter Database. It is a useful resources for such questions.

Bhikkhu Pesala wrote: ↑Tue Feb 20, 2018 6:48 pm

Alfred wrote: ↑Tue Feb 20, 2018 6:43 pmI'd be interested to know if there are any other languages that use it.

Bookmark the site that I referenced. The Letter Database. It is a useful resources for such questions.

Dotless i Usage.png

Thanks, Bhikkhu Pesala. It wasn't obvious that each character in the language table is linked to a page with the relevant character data.

http://www.eki.ee/letter/chardata.cgi?ucode=0131

The Latin alphabet written in the Carolingian script has a dotless small letter I. Indeed, all scripts before the Carolingian had a dotless I- Uncial, Rustic, Half-Uncial and all the others up to the Carolingian, Beneventan and Lombardic use dotless i. The dotless I started to appear with some Gothic scripts in 1200 AD and became stablished with the Humanist script invented by Petrarch.

All my fonts are after calligraphic scripts, including the Trajan script, the most difficult. So the dotless I is a requirement for some languages but also for our Roman scripts pre-humanistic.

The Turkish language adopted the Roman alphabet hence the 26 letters plus a lot more diacritics. The dot on the I is an atonic diacritic, different from the stressed diacritics of the grave and acute diacritic more often referred to as accents. Not to mention those multiple diacritics of ancient Greek, the Greek of Homer, Sappho, etc.

No harm done but it surprised me to find this dependence as the only exception. Regarding the generation of composites the method you implied is fine but not exclusive. It may be suitable for some but not for others.
Incidentally, all modern Latin fonts are essentially based on the Humanist script.

Thanks for the education!