Font View

The font view is a list of all the characters available in the given font.

You may chose to display it at various sizes, by default it will be displayed with the outline font rasterized on a 24 pixel em square. You may display it at 36 and 48 pixel sizes as well. You may also choose to look at an anti-aliased greymap (the above image is anti-aliased). These are slower to generate but look nicer (Comparison).

If you have an encoding which has nothing in it (as opposed to an encoding which maps to a space glyph) then that will be shown by a faint red X drawn across the box.

If you look at a bitmap version of the font, and it does not contain a character for which there is an outline character (ie. the font needs its bitmaps regenerated) then the missing character will be outlined in red to bring it to your attention more easily.

If one of your characters has something in its background layer then this will be indicated visually (currently by darkening the background of the character label).

Small bitmap fonts will be magnified. If the pixelsize of the font is less than 10 it will be shown 3 times normal size, if less than 20 twice normal size.

Underneath the menu bar is some information on the last character selected. First is the encoding of the character in the current font expressed in decimal. Next is the unicode code point for the character (or ???? if it is not a unicode character). Next the postscript name for the character, and finally the unicode name. Not all unicode characters will have unicode names (the CJK characters do not for example)

You may also display any of the bitmap fonts you have generated from the outline characters. Simply go to the View menu and select the one you want.

You may display metrics lines (baseline, origin and advance width) in the font view. (View -> Show H Metrics). If you do so, I suggest viewing the font at a large pixel size, otherwise the window looks too confusing.

A 96 pixel display showing the various horizontal metric lines.
The blue line is the baseline
The small red tick on the left marks the character origin.
The green line on the right shows where the advance width is
 while the bottom green line shows how long it is.
 (normally you will not display both at once)
The character is centered horizontally, and the font ascent is
the top of the box displaying it, while the descent is the bottom.
A 96 pixel display showing the various vertical metric lines.
The blue line down the center is the vertical baseline.
The red bar at the top (over writing the ascent line) is the
 vertical origin.
The green line at the bottom shows where the vertical advance
 is while the green line on the right shows how long it is.
 (normally you will not display both at once)

You may select any number of characters in the font by dragging through them. You may use shift-click to select (or deselect) additional characters. You may also use the arrow keys to move around (and shift-arrow to select). To some extent the order you select characters in will be remembered and a few commands will make use of that.

You may drag and drop selected characters into

The metrics view
Where they will be inserted into the display before the selected charater (selected in the metrics view) in the order in which you selected them in the font view.
The outline view
Where they will appear as references
The char info substitutions/ligature dlg
Where they will appear as a substitution or ligature.

The [Tab] key will move to the next character with something in it, and Shift-[Tab] will move to the previous character with something interesting.

You can perform various operations on the selected characters:

And, of course, you can cut and paste. Normally copying a character will copy that character and any bitmaps associated with it, but you can use the Edit->Copy From menu item to change this so that only entries from the currently displayed font are copied. So if a 12 pixel font is being displayed then only the bitmaps from the 12 pixel bitmap font will be copied in this mode. Cut and Clear will clear those things which would be copied.

Paste is a little more complicated. In general it will paste whatever is in the clipboard, regardless of the copy mode. So if the clipboard contains a 12pixel bitmap and the font view is displaying outlines, then pasting it will paste it into the 12pixel font none the less. There is one exception to this. If the clipboard contains a 12 pixel bitmap and the font view is currently display a 17pixel bitmap then the paste will go into the displayed bitmap.

A Paste from the font view will delete whatever splines were in the character beforehand (while a paste in the outline character view will merge the new set of splines with the old). Paste from the font view will also set the width of the character.

You can scroll the display to any particular character by typing that character. There is also a dialog which allows a slightly more general way of jumping around. You may type in the character's name (all characters in a font will be named), its encoding value, its unicode encoding value, or (for 2 byte fonts) its ku ten specification.

Pressing the right mouse button invokes a popup menu.

Encodings and Character Sets

A Character Set is a set of characters (for example all the letters of the alphabet would be one character set). An encoding is a function that takes a number (often between 0 and 255) and maps it to a character. Often these two are used synonymously as an encoding generally implies a given character set. (strictly speaking a character set is an unordered collection with no implied encoding, and some encodings work with multiple character sets).

All fonts in pfaedit have both a character set and an encoding. The character set is just the set of characters in the font. The encoding is the way those characters are ordered (or sometimes, the way the first 256 are ordered). The font info dialog allows you to chose the encoding (and character set) you wish for a font. Often a font will have a few extra characters that don't fit into the character set specified by the encoding, these characters will be placed at the end of the font and when the font is output (ie. postscript is generated, or a bitmap font created) the characters will be included in the font but they will not be encoded (this can be useful, especially in postscript where it is possible to reencode a font at run time. Thus a font might have all the characters needed both for a cyrillic character set and for a latin one, but only one character set at a time would be encoded).

It is possible for an encoding not to be one-to-one. That is there may be several characters that might lead to the same underlying glyph (the classic example is the non-breaking space which often uses the same glyph as the breaking space). When PfaEdit encounters such an encoding it will create a character structure for each entry, one of those structures will contain the data for drawing the glyph, the other(s) will contain a reference to the one real character. All characters will have the same name. When PfaEdit generates a font it looks for this case and turns it into a font whose encoding vector contains multiple references to a glyph.

You may also create a custom encoding. (Generally, only the first 256 entries of a Custom encodings will actually be part of a font's encoding vector). You may change the name of any character in the font with the Element->Char Info command. This will force the font to have a custom encoding.

There are many standard encodings built in to the program. There are 14 ISO 8859 encodings, the encoding used by the Macintosh for the US and Western Europe, the encoding used by MS Windows for the US and Europe (which is a slight extension of ISO 8859-1). There are also several 2-byte encodings built in. Several of these are subsets of unicode whose goal is to specify every character currently used by humans. There are 4 CJK (Chinese, Japanese, Korean) 94x94 encodings (two bytes where each byte ranges over the 94 ASCII printable characters excluding the space character). And there are sjis (Japanese), big5 (Traditional Chinese), Wansung (Korean) and johab (Korean) . You can also add your own encodings with the Element->Font Info dialog. A brief description of what's in the encodings also appears on that page.

There may be some slight differences between one vendor's definition of an encoding and another. I try to follow those encoding specified by the Unicode consortium. I notice that the Symbol encoding used by Microsoft differs from the Unicode one (usually just be substituting synonyms, Omega for Ohm sign for instance).

Adobe did not choose their naming conventions very well for Greek letters. They assigned the name "mu" to Micro Sign, "Delta" to Increment and "Omega" to Ohm Sign. So the Greek alphabet has ugly nameless holes in it. I have cheated by reassigning "Delta" and "Omega" so that they refer to the letters, but "mu" is part of Adobe's Latin1 encoding and I can't change that.

CID fonts (see below) have no encodings. Instead they are designed to be associated with one or several cmap files which provide encodings in a general way. cmap files are beyond the scope of PfaEdit. Adobe has defined many which are freely available.

There is one final encoding called "Compacted". This encoding is just a contiguous list of glyphs with no empty slots in it. It is often easier to look at a CJK or unicode font without having to wade through thousands of unused code points.

Unicode (ISO 10646)

PostScript® assigns a name to every unicode character. Some of these names are fairly obvious like "A" for the first letter of the latin alphabet, others are more obscure like "afii57664" for hebrew alef, while others are just "uni8765" for the unicode character at 0x8765.

PostScript Unicode encoding does not quite mesh with the unicode standard. This is probably because PostScript deals in glyphs and Unicode deals with characters so PostScript sees no distinction between space and nobreak space (and so does not encode the latter) while Unicode does.

According to the unicode website the first 256 character positions of unicode and ISO 8859-1 (ISO Latin1) are the same. However the PostScript encoding of ISOLatin1 is slightly different from that specified for unicode. I cannot explain the reasoning behind this. PfaEdit does not use PostScript's ISOLatin1 Encoding vector, instead it uses the first 256 code positions of the unicode encoding vector (slightly modified by me to conform to Unicode conventions rather than PostScript. So I include nobreak space and the soft-hyphen).

PfaEdit supports two slightly different unicode encodings. The first contains only the first 65536 characters of unicode (those in the Basic Multilingual Plane, or BMP), while the second can contain as many characters as you have memory for. Very few characters have been defined outside of the BMP so at the moment you probably want to use just the BMP (the full encoding isn't completely supported yet anyway). Currently there are only definitions for plane 0 (BMP, U+0000-U+ffff), Plane 1 (SMP, Secondary Multilingual Plane, U+10000-U+1ffff), Plane 2 (SIP, Supplementary Ideographic Plane, U+20000-U+2ffff), and Plane 14 (SSP Supplementary Special-purpose Plane, U+e0000-U+effff).

CID keyed fonts

The standard mechanisms that postscript provides work reasonably well for alphabets and syllibaries, but the massive number of characters needed for CJK (Chinese, Japanese or Korean) fonts require more complex machinations. Adobe's current solution is the CID-keyed font, a font consisting of several subfonts each a collection of glyph descriptions with no encoding imposed on them and no names given to them.

If PfaEdit loads in a CID keyed font it will enable a special menu called CID which (among other things) displays the list of all the sub-fonts in the font. When it starts up PfaEdit will pick (practically at random) a sub-font to display in the font view. You may change which font is displayed by the CID menu.

The CID menu also allows you to turn a normal font into a CID keyed font. It creates a collection containing just the original font (you may add other fonts, or blank fonts, later). But before it can create a collection PfaEdit needs to know what character set you will be using. A character set is just a collection of characters, and you may define your own if you wish (but if you do you have to define your own cmap files, etc and it's probably not worth it). Adobe has defined character sets for Japanese (actually there are two, one corresponding to JIS208 and one to JIS212), Korean (Wansung & Johab), Traditional Chinese (Big5) and Simplified Chinese (GB2312). Although these character sets are based on the standards mentioned, they each have many additional characters. Adobe also defines a charset that works for Unicode, but it is called "Identity" instead. PfaEdit does not have these character sets built in to it, each must be loaded the first time it is used. I provide one file for each of the above character sets, they have an extension of ".cidmap", and you may download them all from here.

(There are also a few other character sets floating around but they aren't in the public domain and I haven't been able to get my hands on them).

Adobe identifies each character set by a three values: a registry, an ordering and a supplement. The registry is the name of the organization that is defining the character set (Adobe calls itself Adobe), an ordering identifies the character set (Japan1, Korea1, etc.) and the supplement indicates how many times the ordering has been revised. Characters may only be added to an ordering, never removed, so an old font will be perfectly described by a newer character set, while a new font described by an old character set will have some unavailable characters. So a full cidmap name will look like:

Warning: CID keyed fonts (actually any CJK font) use massive amounts of memory in PfaEdit. I am able to work on many of them on my 256M machine with 384M of swap space.

The Remove Undoes command will allow you to free up memory if you think you may be running short. PfaEdit is not always able to protect itself against running out of memory, sometimes the OS just sends it a SIGKILL signal.

Multiple Master Fonts

If the font is a Multiple Master Font there will again be several subfonts only this time all the subfonts contain the same character set. Each subfont provides glyphs for one style of the font family. The MM menu allows you to control which style of the family is visible in the font view, and provides a few other commands for manipulating multiple masters.

Vertical Metrics

CJK fonts generally should have vertical metrics. Latin (Cyrillic, Greek) fonts generally should not. If PfaEdit reads in a font with vertical metrics it will retain those metrics, but when PfaEdit creates a new font then that font will not have vertical metrics enabled.

To enable vertical metrics for a font go to Element->Font Info and select the General tab and check the [*] Has Vertical Metrics checkbox. This will allow you to set the vertical origin for the font and it will give every character a default vertical advance of the emsize of the font (ascent+descent).

The vertical origin is the y height (in the design coordinate system) of the origin for vertical metrics.

In the outline character view you will be able to adjust the vertical metrics just as you adjust the horizontal metrics.

You can view the vertical metrics in the font view just as you can the horizontal metrics with View->Show V Metrics.

-- Prev -- TOC -- Next --