On Friday 28 Jan 2005 20:50 pm, Ted Hilts wrote:

So on Linux just how does one go about creating a new font?

Well, on the 9.1 install media there are xmbdfed for editing 'bdf' fonts and PfaEdit for PostScript fonts. I've never used either though. There should be other packages available if you google for them.

and So on Linux just how does one install a font?

That depends. What desktop are you using and what SuSE version?

and Is there an international standard for language fonts?

Yes, more than one. Most common IME are ISO-8859 and unicode (www.unicode.org)

and What are the tools (on the desktop) for creating and for installing fonts?

Hope these are not dumb questions. I do know one very small thing. I know there are files like type "ttf" which contain a font (or set of font characters like A, B, C, etc.?) and the data in the file is according to some format of information which completely defines the font.

Each different font format stores the data in different ways. The simplest is a BMP (=bitmap) font which literally stores a fixed bitmap of each character. Truetype fonts store a set of points and curve segments for each character relative to a base grid, and information about spacing, direction, etc. which allows them to be dynamically scaled and manipulated. As to other formats, I couldn't say, for sure, but reasonable possibilities include a set of primitives (curves, loops, risers and descenders etc...) and instructions for combining them to create each character; decomposing characters into geometric primitives (lines, triangles, etc...); vector-graphic type descriptions. Of course, different encoding formats lend themselves to different applications and different character sets. BPM fonts are fast, and very suitable for text consoles. Truetype fonts are reasonable for text manipulation up to a point. For alphabetic writing systems like English, separate encoding of each character is acceptable since there is a fairly small inventory. Some alphabetic systems (like Devanagari) combine regularly occurring adjacent letters into compound forms, so the sequence of printed characters doesn't necessarily match the number of 'letters' in the string. In some cases this can lead to hundreds of distinct symbols. Syllabic systems (like Japanese Katakana) often run up to 80 to 100 characters, but are also often 'built' from a base symbol and a modifier. More complicated from a font point of view, but actually much easier to read and write for native users, are ideographic or logographic systems like Chinese - in principle every word can have its own character, although in practise that is rarely true. Chinese has at least 50000 identifiable characters, and a fluent reader is likely to know 15-20000 without need of thought or dictionary. Lastly, there are pictographic systems, Egyptian Heieroglyphs being the best known, but Mayan being the most complex as the glyphs for a single word are not written as a string or in a cartouche but morphed into a regular square.

That's not much to go on. I don't even know if each character is a font or a set of characters constitute a font.

Usually, the font file will contain all the characters of the font, although a quick look through the basic ones will show that many have gaps. A font is the complete set of type for a particular typeface. For a printer this would include various sizes and common variations like italics or bold. Sometimes, there is one file for the font, and the rendering engine creates italics and bold, as well as scaling. High quality fonts, however, often have separate files containing italic, bold or bold-italic variants to maintain the print quality, and they may also have various sets of files for different scale ranges. Dylan

Ted

-- "I see your Schwartz is as big as mine"                                   -Dark Helmet