Sourcing Composites

[outofstepfontco]

I am getting a lot of use out of Complete Composites but I struggle with some characters.

I see there is a "Used By" feature that lists characters that use the simple character to complete a composite.

Is there any way to source a composite to find out what simple characters are used to build it? Similar to the auditing function in Excel. I find my I and G characters are often blank and I can't figure out why. It would also help keep me wasting time building characters that are built from composites.

Haven't been able to find any information in the help files or searching the forums - your help is much appreciated.

[Bhikkhu Pesala]

Double-click on a composite glyph member in the glyph edit window to open the Composite Glyph Properties dialogue.

Composite Glyph Properties.png (48.83 KiB) Viewed 6905 times

Right-click on a composite glyph member to open it in a new glyph edit window, or to jump to it in the current window.

Open in New Window.png (7.65 KiB) Viewed 6905 times

[Bhikkhu Pesala]

The Complete Composites Tutorial thread will tell you more, and the PDF file linked to at the end of the thread goes into the technical details of how it works.

If some of your composites are blank or incomplete, probably the diacritics used to compose them are missing or empty.

Some use a combining overline, which will be absent for most users, but there should be a fallback glyph (the underscore character). If that's also missing, then add it.

For exampe: g stroke is composed of lower case g + combing long stroke overlay (decimal code-point 822). If the font doesn't contain that glyph, the underscore glyph is used instead (decimal code-point 95)

Code: Select all

<Composite><!-- g stroke -->
<GlyphMapping>485</GlyphMapping>
	<Member id="1">
	<GlyphMapping>103</GlyphMapping>
		<UseMetrics>TRUE</UseMetrics>
	</Member>
	<Member id="2">
	<GlyphMapping>822</GlyphMapping>
		<GlyphMapping>95</GlyphMapping>
	</Member>

[Bhikkhu Pesala]

After using Complete Composites to compose the g stroke glyph, it looks like this:

g stroke.png (4.83 KiB) Viewed 6904 times

Then, one may need to make it simple, and make some adjustments before combining the two glyph members using Get Union of Contours. Then it looks like this:

g stroke completed.png (4.68 KiB) Viewed 6904 times

[outofstepfontco]

Double-click on a composite glyph member in the glyph edit window to open the Composite Glyph Properties dialogue.

Composite Glyph Properties.png

Right-click on a composite glyph member to open it in a new glyph edit window, or to jump to it in the current window.

Open in New Window.png

The Open New Window function only provides me with the source if the simple glyph is in the composite. For example: $0135 J circumflex: Blank after Complete Composites. Why when I have a small J. There is nothing for me to click to open in new window.

[outofstepfontco]

If some of your composites are blank or incomplete, probably the diacritics used to compose them are missing or empty.

Absolutely the diacritics are empty. There is no doubt.

But $0135 small j circumflex is empty after complete composites, and $006A small J doesn't even list $0135 in its Used By list.

So how do I find out what characters I need to fill in order to get $0135? And how do I use the method to solve my other similar problems?

It sounds like the software needs the function added.

[Bhikkhu Pesala]

For example: $0135 J circumflex: Blank after Complete Composites. Why when I have a small J. There is nothing for me to click to open in new window.

Composites with j require dotless j. Thanks for reporting the issue. The composite definition. should perhaps have a lowercase "j" glyph as a fallback. Composites with lowercase "i" suffer from the same issue. They require a dotless ı to be present.

I will show you how to modify CompositeData.xml to fix that, but it would be much better to fix the font by adding those two glyphs.

On second thoughts, I am not sure if it's even a good idea to use lowercase i to compose í ì ï ǐ etc. Users would only have to make them simple, and delete the dot above the i (or j). It is better that they learn how to design the font correctly.

If you were to do it by editing CompositeData.xml, you would add a fallback glyph after the dotless i (305)

Code: Select all

<Composite><!-- i grave -->
<GlyphMapping>236</GlyphMapping>
	<Member id="1">
	<GlyphMapping>305</GlyphMapping>
	<GlyphMapping>105</GlyphMapping>
		<UseMetrics>TRUE</UseMetrics>
	</Member>
	<Member id="2">
	<GlyphMapping>57365</GlyphMapping>
	<GlyphMapping>96</GlyphMapping>
		<Pos>Auto</Pos>
	</Member>
	</Composite>

The fallback glyph of (96) for the low profile acute accent (57365) is there, and necessary. Most users won't have that glyph in their fonts, and its not essential. If you want to use smaller accents for narrow letters like î then it's useful to have. They are also useful for Uppercase accented glyphs. Read the PDF for details.

[outofstepfontco]

The lowercase j is not the problem.

The problem is that when glyphs show up blank, there's no function that tells us why.

I took the XML file you gave me and stripped out the data to provide the attached table. It's a Codepoint (Unicode) list of all the glyphs that your program creates using composites. This is what I needed for direction on which glyphs I don't need to build when I'm building glyphs. For example, I didn't know that I didn't need to create my own colon, semicolon, feminine ordinal, masculine ordinal, or cent. But now I have a reference list when working in my other program.

I could probably build a table that also states the source for each of the composite glyphs, but that would take me more time than I have today. Also, you might already have such a that you could post, so it might be a waste of my time.

[Bhikkhu Pesala]

The problem is that when glyphs show up blank, there's no function that tells us why.

It could be made more user-friendly, but I think it should be obvious that a composite glyph requires the glyph members to be present.

Anyone who doesn't know what a composite glyph is might well be confused why Complete Composites shows only the base glyph, or only the accent, or nothing at all, but anyone who does know what a composite glyph is should soon fathom it out. If not, that's what the support forum is for. The manuals cannot possible include everything, and if they did, no one would read it all.

For glyphs like colon, semicolon, Spanish Ordinals, Euro sign, cent, etc., Complete Composites is no more than a design aid. Although you could leave semicolon as a composite, cent, Euro, and Spanish Ordinals will definitely need some work to make them look right for the font, and many users may prefer to start from scratch. The main benefit for the currency symbols is that they use the figure width (zero), which is good design for tabular figures.

[outofstepfontco]

Anyone who doesn't know what a composite glyph is might well be confused why Complete Composites shows only the base glyph, or only the accent, or nothing at all, but anyone who does know what a composite glyph should soon fathom it out. If not, that's what the support forum is for. The manuals cannot possible include everything, and if they did, no one would read it all.

At the same time, I demonstrated significant knowledge of what Composite glyphs are, and how they are formed, but I didn't know the $0135 small j circumflex requires a dotless j U+0237 until you said so in your forum response.

That is a problem for all users, not just for me.

This is not the only character that this happens for. There are i glyphs, g glyphs, a huge range of punctuation glyphs, combining diacriticals, spacing modifier characters. The only way to know why a glyph is blank is to memorize its source. This is a completely unnecessary learning curve.

[Bhikkhu Pesala]

That is a problem for all users, not just for me.

So far, in the ten years or so since the Complete Composites feature was added, you're the first one to report the problem, otherwise I would have known about the issue, and perhaps Erwin would have done something to find a better solution.

I have reported the issue to Erwin, so it's definitely an area that could be improved, but as far as I can tell, it has not been a problem for anyone until now.

Open any standard Windows font, one like Calibri that was not even made with FontCreator, and see how the i circumflex, i grave etc., are composed. They are not composed from lowercase i.

I agree that someone new to font development won't know how fonts are made, and FontCreator should try to make life as easy as possible. Even for experienced users it would be helpful to remind them, "You haven't added a dotless i to your font yet."

[outofstepfontco]

Okay, duly noted.

Right now I'm working on a font in my modular font-building program in one window, and I have Excel open in another window showing the Name + Unicode for all the composites that FontCreator will build for me. I see that Plus Sign, Minus Sign, Dot Accent, Ogonek, but no Small Tilde ($02DC). So I fill in the Small Tilde and keep going down the list. I keep going down the list, skip Rhotal, and I see all the Low Down Arrowheads are missing from the composites list, $02EF to $02F2. Better fill them in.

I believe in simplifying workflows, and am happy to make my resources available to other users who are interested in the same.

[outofstepfontco]

Hi everyone.
I've processed the Composites XML file and done my best to pull out the source glyphs for each composite glyph.

There are too many glyphs for me to check, so I can't guarantee that all the source glyphs in the list are correct, but it's a start.

Hope this helps speed up your workflow.