delahunt@adobe.COM (Jim DeLaHunt) (10/28/89)
In article <2362@eagle.wesleyan.edu> rsilverman@eagle.wesleyan.edu writes: >In article <5203@ogccse.ogc.edu>, goward@ogccse.ogc.edu (Philip Goward) >writes: >> I've been wondering how PostScript adapts to 16-bit fonts, such >> as Kanji. How are characters with codes > 255 accessed? ... > >The standard definition of PostScript does not allow for more than 256 >characters to be encoded at once .... I remember reading that sometime >in the past few months, Adobe started marketing an extended version that >could deal better with the requirements of non-Roman writing systems. Actually, "16-bit characters" is an oversimplification, and Adobe's OEMs have been selling printers with an extended PostScript(R) interpreter and kanji fonts for two years..., but I should give a complete overview. You can indeed print Japanese with PostScript(R) devices, and there are currently three laser printers (DEC, NEC, Apple) on the market, with two typesetters (Varityper and Linotronic) to join them soon. Support for Japanese (and Chinese and Korean) involves two parts: extensions to the PostScript language (i.e. an upgraded interpreter), and the appropriate fonts. The three main technical challenges of these languages are: * Large numbers of characters (over 7,000 for Japanese) * Multiple writing directions (horizontal and vertical) * Many inconsistent standards for character set and byte encodings The most important language extension is the "composite font". It is a font which, instead of containing character descriptions, contains fonts -- in a hierarchy of fonts. At the bottom are "base fonts", identical to the roman fonts we all know, which contain character descriptions. Bytes from a show string select a base font from the composite font, and then a character from the base font (and the font can select from number of algorithms for doing this). Also, fonts (composite or base) can now contain two sets of metrics; a key in the font selects one for the 'show' operator to use. In kanji fonts, one set of metrics moves the current point horizontally, and the other vertically. There are a number of other minor extensions. The extended interpreters are completely backwards-compatible; they support Red Book PostScript and roman fonts fully. The hierarchical nature of the composite font strategy is helpful in many ways. A two-level hierarchy can address tens of thousands of characters, and current interpreters support up to a five level hierarchy. Japanese vertical and horizontal fonts are almost, but not entirely, identical; so the two top-level composite font dictionaries (which are small) may share, say, 78 out of 80 base fonts (which are not small). And character set standards, while differing, overlap considerably, and again can share base fonts. Adobe and the Japanese type foundry Morisawa have developed two Japanese typefaces, Ryumin-Light and GothicBBB-Medium, with more to come. They include the complete JIS-83 and JIS-78 character sets, plus some extras. These typefaces are currently available only on printers, but when more typefaces appear they will be available as aftermarket (downloadable) fonts. The reason that "16-bit characters" is an over-simplification is that it doesn't reflect the diversity of encodings and character sets. "JIS-83" or "the NEC PC character set" are better. Besides, Japan will soon adopt a new character set standard (they intended last year to call it "JIS-88" :-) ), which will involve 3-byte encodings. The composite fonts extensions already support 3-byte encodings. This has been rather long, but I hope informative. You can get documentation from the Adobe file server (send mail containing the word "help" to ps-file-server@adobe.com). -- --Jim DeLaHunt +1-415-962-3790 delahunt@adobe.com (Internet) PS Technical Support ...!{sun,apple,decwrl}!adobe!delahunt (UUCP) | Adobe Sytems, Inc. | P.O. Box 7900 | Mountain View, CA 94039-9923 | USA | The above are my personal opinions, and do not necessarily reflect Adobe's.