Encoding Menu

The Encoding menu is only found in the font view.

Reencode

Has an attached sub-menu of standard and user defined encodings. The font’s current encoding will be indicated with a check mark. You may change the encoding by selecting a different entry.

There are two slightly different formats an encoding can take. It can be defined by unicode code points, or it can be defined by glyph names. When reencoding to an encoding defined by code points, the glyph with the matching unicode value is placed in the encoding slot. When reencoding to an encoding defined by glyph name, we first search for a glyph with the matching name and use it, if not found we search for the glyph name’s corresponding unicode code point (if any) and if found we change its name to that specified by the encoding.

Example: Suppose we have a font containing a glyph named “uni0041”, and an encoding which maps U+0041->slot 65, then the glyph will be moved into slot 65. If we have another encoding which maps “A” -> slot 65, then (since “A” has unicode value U+0041) our glyph will still be mapped to slot 65, but in addition its name will be changed to “A”.

Compact

Remove any holes from the encoding so all the glyphs get smushed together. If the font is already compact, then selecting this again will restore the original.

Force Encoding

Has the same sub-menu as above. Here we assume that the glyphs of the font are currently encoded in the right order, but they have the wrong names (This may seem odd, but it happens a lot). This command will change the names of all the glyphs to match what they should be if the indicated encoding were in force.

Add Encoding Slots...

Add some extra slots at the end of the font into which you can put unencoded glyphs (variant glyphs, etc.)

Remove Unused Slots

Removes any unused slots from the end of the font. It does not remove unused slots inside the font, that would screw up the encoding.

Detach Glyphs

Detaches any selected encoding slots from their currently associated glyphs. These slots will now be marked as unused. The glyphs will remain in the font, just not encoded (If you reencode the font those glyphs will become visible again).

Detach & Remove Glyphs...

Similar to the above except that any glyphs detached (which are not used elsewhere in the encoding) will be removed from the font.

Add Encoding Name...

Requests an encoding name from the user and searches for it in the iconv() database. It then adds that encoding to the menu.

Load Encoding...

Asks the user for a filename and attempts to load a user defined encoding from that file. (You can only load small encodings – one byte encodings)

Make from Font...

Allows you to name the font’s current encoding (if it isn’t already named), and add it to the encoding menu.

Remove Encoding...

Removes one of the user defined encodings from the menu.

Display by Group...

Allows you restrict the glyphs displayed in the font view to those in a user defined group (specified in the next command).

Define Groups...

Allows you to define groups of glyphs which (presumably) have some meaningful connection to each other.

Save NameList of Font...

Creates a namelist file mapping unicode to the glyph names of the current font.

Load NameList...

Loads a namelist file into fontforge and copies it so that fontforge will load it on start up in the future.

Rename Glyphs...

Allows you to specify a namelist. All glyphs in the current font will be renamed to match the scheme in the namelist.

Create Named Glyphs...

Allows you to specify a file containing a list of glyph names. FontForge will create a sequence of unencoded glyphs with these names. This might be useful if all your fonts contain small caps and you always want to have the names “A.small”, “B.small”, “C.small” etc. or perhaps you always want the ligatures “longs_longs_t”, “f_longs”, “f_j”, etc.

General notes on encodings

Not all font formats support all encodings. SVG fonts will always be output in a unicode encoding, truetype fonts in either unicode or one of the CJK encodings, type1 fonts only support single byte encodings, etc.

In a CID keyed font you are not allowed to change the encoding (in essence because there is none), but there is an entry CID->Change Supplement which will display the Registry/Ordering information and allow you to change the supplement.

Built in Encodings

FontForge knows about the following encodings by default:

  • ISO-8859-1 (Latin1) – traditional encoding for western european characters. Default encoding for http. Does not include the Euro sign.

  • ISO-8859-15 (Latin0) – Replacement for Latin1. Does include the Euro.

  • ISO-8859-2 (Latin2) – Central & Eastern European (Czech, Hungarian, Polish, Romanian, Croatian, Slovak, Slovnian.

  • ISO-8859-3 (Latin3) – Southern European (Esperanto, Maltese)

  • ISO-8859-4 (Latin4) – Northern European (Estonian, Latvian, Lithuanian, Greenlandic, Lappish)

  • ISO-8859-9 (Latin5) – Turkish

  • ISO-8859-10 (Latin6) – Nordic (reworking of Latin4&Latin1)

  • ISO-8859-13 (Latin7) – Another Baltic character set

  • ISO-8859-14 (Latin8) – Celtic (Gaelic & Welsh)

  • ISO-8859-5 (Cyrillic)

  • ISO-8859-6 (Arabic)

  • ISO-8859-7 (Greek)

  • ISO-8859-8 (Hebrew) – (and Yiddish)

  • ISO-8859-11 (Thai) – Also know as TIS 620

    <there is no ISO-8859-12>

  • KOI8-R – Cyrillic

  • Macintosh Roman

  • Windows “ANSI” (CodePage1252)

  • Adobe Standard

  • Symbol

  • TeX Base

  • ISO-10646-1 (Unicode, BMP)

  • ISO-10646-1 (Unicode, Full)

  • ISO-10646-? (Unicode, by plane)

    (You can select a specific plane of unicode as an encoding (ie BMP, SMP, SIP,…)

  • SJIS

  • JIS 208 – Japanese Kanji (first 8000 characters)

  • JIS 212 – Japanese Kanji (next 8000 characters)

  • Wansung

  • KSC 5601 – Korean (this is the 94x94 version of KSC 5601)

  • Johab

  • GB 2312 – Simplified Chinese

  • Packed GB 2312 – (I don’t know what the proper name for this is, ASCII for bytes<0x80, and GB 2312 EUC offset by 0x8080)

  • Big5 – Traditional Chinese

  • Custom – An unknown encoding

  • Glyph Order – the glyph ordering used in the original font file.

Encoding sources:

An index to images of all the glyphs in unicode.

User Defined Encodings

You can also add new encodings to the set that FontForge knows about. There are three menu items that manipulate a set of user defined encodings. As always these specify both a character set and an encoding. The encoding has a maximum of 256 entries, but the character set may be larger (up to 1024). This means that you can define a font with extra characters. Since postscript fonts can be reencoded at runtime this can be useful.

The Load Encoding command allows you to load an encoding(s) from a file. Currently the file must either be in the format used by the unicode consortium for mapping ISO 8859 encodings to unicode, or it must be a postscript encoding array. The first format looks like this:

0x20   0x0020  #       SPACE
0x21    0x0021  #       EXCLAMATION MARK
...

A postscript file looks like:

/TeXBase1Encoding [
 % 00
 /.notdef /dotaccent /fi /fl
 /fraction /hungarianumlaut /Lslash /lslash
 ...
] def

There may be more than one encoding in a postscript file. The encoding parser is not smart. It will only read arrays specified like this, don’t try any of the innumerable other ways of specifying an array in postscript.

If the font has a custom encoding then the Make From Font menu item is enabled. This allows you to name the encoding you have defined for the current font.

The Remove Encoding menu item brings up a list showing all the custom encodings and allows you to delete them.

Here’s an example of a postscript encoding file. It contains:

  • TeXMathItalicEncoding

  • TeXMathSymbolEncoding

  • TeXMathExtensionEncoding

  • IsoLatin – (which specifies all the characters used in any of the ISO-Latin-* fonts

  • AdobeExpert – (Which contains things like lower case numbers, small caps, fractions, sub/superscript numbers, etc.)

  • CodePage1250 – Microsoft’s encoding for Central European characters

  • CodePage1251 – Microsoft’s Cyrillic encoding

  • CodePage1252 – Microsoft’s Western European encoding (a superset of Latin1. Sometimes called “ANSI” though I can find no ANSI standard that it follows)

  • CodePage1253 – Microsoft’s Greek encoding

  • CodePage1254 – Microsoft’s Turkish encoding

  • CodePage1255 – Microsoft’s Hebrew encoding (an extension of ISO-8859-8)

  • CodePage1256 – Microsoft’s Arabic encoding

  • CodePage1257 – Microsoft’s Baltic encoding

  • CodePage1258 – Microsoft’s Viet Namese encoding

  • CodePage874 – Microsoft’s Thai encoding

  • MacCentralEuropean

  • MacCyrillic

  • MacGreek

  • MacHebrew

  • US-ASCII – Not really useful by itself any more, but provides the first 128 characters of almost every other encoding.

NameLists

Adobe has established a standard glyph naming convention which provides intelligible names for many glyphs of unicode characters. And some unintelligible names too.

A namelist is just a mapping from unicode to glyph names (a glyph name must be made up of alphanumeric characters (or the special characters ‘.’ or ‘_’), it may not begin with a digit, and it must be 31 or fewer characters in length.

FontForge provides a series of standard namelists:

  • Adobe Glyph List

    This is the set of names that Adobe publishes on the web.

  • AGL For New Fonts

    This is set of the names that Adobe recommends for the new fonts, without odd names like “afiiXXXXX” or incorrect commaaccent names.

  • AGL without afii

    The cyrillic and hebrew glyphs have been assigned some very odd names (afiiXXXXX) and some people prefer not to use them. This is now redundant; AGL for New Fonts should be used instead.

  • AGL with PUA

    Adobe has assigned part of the unicode public use area to hold some standard glyph variants like small caps, subscripts, old-style numbers, etc.

  • Greek small caps

    I’ve added some greek small cap assignments

  • TeX Names

    The TeX typesetting system has its own set of names

  • AMS Names

    The American Mathematical Society has its set of names (see the American Mathematical Society’s specification)

You may define your own namelist file. It should have the extension “.nam”. And it should contain a series of lines that look like:

0x0020 space
0x0021 exclam
0x0022 dblquote

In many cases you will just want to make a few modifications to an already existing namelist. You can write:

Based: Adobe Glyph List
0x0021 exclamation

Which means the namelist is the same as the Adobe Glyph List except that in your system U+0021 will be called “exclamation” rather than “exclam”.

Any glyphs you do not explicitly name will be named “uniXXXX” or “uXXXXX” where XXXX is the unicode value in hex. The prefix “uni” will be used for glyphs in the BMP, the prefix “u” for glyphs outside.

I said that glyph names should contain only alphanumeric characters (and some others). But really this is only true of glyph names that get output into a font. If you design your own namelist you may include whatever glyph names you like, using (almost) the full range of unicode – provided you rename all your glyphs to Adobe’s names before you generate a font. If you use non-ASCII characters you should encode the file in utf-8.

FontForge normally will restrict glyph names to be within the ASCII range, but if you go to File ‣ Preferences ‣ Generic ‣ UnicodeGlyphNames and set it on you can use unicode.

I’ve created a french namelist which uses accented letters. This will work fine within fontforge, and most rasterizers will parse fonts generated using such names – but they are non-standard and may cause problems. Best to do a rename just before generating your font.

So if you find it easier to work with names other than those Adobe has established you may create your own namelist file. Then use Encoding->Load NameList… to load that into FontForge (you only need do this once, FontForge should remember it thereafter). You may use File->Preferences->Font Info to decree that all your new fonts will use this namelist. You can change a font’s namelist with either:

You may also want to force a rename of all glyphs when you Open a font, and the open dialog now lets you do this. Similarly when generating a font you will probably want to force that font to be use the standard names of the Adobe Glyph List.