![]() ![]() The problem with this is that the PostScript interpreter has no information (none at all) about the source character set, it only has information about the glyph identifiers in the Type 1 font program dictionary. In 1991 when Warnock announed in the New York Times that search support was to arrive in Carousel, the Adobe Type 1 Specification and the Apple TrueType 1 Specification had already been published in the summer of 1991, and the Apple and Microsoft had already made it clear that the OS market wanted a character encoding model instead of a glyph encoding model.įrom there it went to bits, Adobe pressing for more and more type designers to move to the Type 1 format in order to try as best as possible to maintain control of the glyph identifiers, Apple and Microsoft pressing in the opposite direction where in the intelligent font model character codes and public and font-independent while glyph codes are private and font-dependent.īy April 1992 when FOGRA Forschungsgesellschaft Druck in München held a conference on the direction of development in professional prepress, Adobe in a telecommunications appearance claimed that 16,000 products were already on the market in the Type 1 font program dictionary format and that Carousel was imminent in 1993 (per online article by Chief Editor Kurt Wolf of Deutsche Drucker). This way incompatible character constitutencies could index into the surrogate superset and character information recovered. In 1989 when Warnock announced in the New York Times that the aim was to distribute a document world wide in wholly digital form, the commercial and mathematical challenge was to turn Adobe glyph names into a surrogate character superset. PostScript glyph identifiers were initially regulated by Adobe and Linotype, but whether or not you could get from the glyph identifiers back to the character information was of no importance because the character information was not unified and thus could not in principle serve as support for an internationalised search architecture. Harrington and Buckley in their 1988 to Interpress explain that the design allows several character sets with incompatible constitutencies to be imaged simultaneously.įor the RIP to process and draw glyphs it has to have an identification scheme of sorts, and that scheme could not be a character encoding prior to ISO-IEC 10646-1:1993. The simple set theory at the basis of this is that Interpress and PostScript were both designed as glyph encoding models and not as character encoding models. (Among other things, this iprinciple is stated in the patent for heuristics to reclaim character information which Adobe has licenced from PDFLib for PDF/A.) There is a first principle in information technology: You can automatically audit whether or not a data structure is present, but you cannot automatically audit whether its meaning is correct or incorrect. I'm curious that this is not automatically possible since, even under the old Type1 PS font tables glyphs were named as well as just given a number. ![]() I can use TrueType 2 with glyph substitution from 1994 in OS X today and have any complexity of composition, but I can use no product from Adobe in 1994 today because Adobe product in 1994 used character substitution instead of glyph substitution which plays havoc with search support. If Adobe and a whole lot of PS type libraries (hallo the Linotype Library in Bad Homburg) had followed the ISO and Unicode lead from the start, none of this overwhelming complexity would be confronted today. If you want, there is a guide to doing it with TrueType 1 by Apple's Unicode technical director in the Apple Text section of the website. Note that if you have many characters that map to one glyph as is the case with ligatures, you must remember both to set up the character-glyph mapping and, if you want an OpenType font file that will yield searchable glyphs in Acrobat, to follow the Adobe glyph name conventions that require you to decompose the glyph semantics by writing UTF16 character identifiers into the glyph name itself. Like it or not, but the one and only way is to pick up a copy of the ISO10646 or Unicode books, look up in the default glyph form for each character code, and then go over the font's character-glyph mapping. This was tried in Mac OS 7.5 with the Type 1 Enabler which was unable to address the issue of incorrect glyph names, so if you had garbage in PostScript you got garbage in TrueType 2. However, you also need to add that an automatic and unassisted conversion from a PostScript font program dictionary to any of the flavours of the Unicode character-glyph model will If the font is causing incorrect character to glyph mapping, then you are correct that the incorrect mapping can be repaired by the enduser. ![]() If it is causing you problems it helps to convert it to a Unicode Opentype font using a utility ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |