Peter Kirk suggested: > Presumably the same principles can be applied when you run into a newly > discovered (probably archaic) cuneiform character. Except that for some > reason, Ken, you classified "dynamic cuneiform" as Type VI: Glyph > Description Language. Why can't it be seen as Type V: Ideographic > Character Description, encoded with "pseudo-operator-like" symbols (Dean > proposed just 14 of them)? That would provide a solution for newly > discovered cuneiform as well as for newly discovered Han.
Such a scheme could, of course, also be suggested, although it wasn't what Dean Snyder was proposing. > > I am not suggesting that this model should be used for common cuneiform > characters any more than for common CJK characters; the static model, > which was once agreed on, can stand. I am suggesting it only as a > mechanism for describing newly discovered characters. I doubt it would be cost-effective for cuneiform. The number of additional signs turning up will (most likely) be small, and the benefit in having a set of symbols to attempt a syntactic description of them from pieces would be small. It would be much more effective in most instances to simply write the description of the missing sign in terms of traditional cuneiform names and "SIGN1 TIMES SIGN2" etc. to indicate various types of productive composition for the new combinations which might turn up. And adding such schemes is not without cost. Even if they were a new set of symbols strictly limited to cuneiform sign description, they would set a precedent that somebody else would try to apply in some inappropriate context. Having the Ideographic Description Characters in Unicode has mostly been a cost -- a cost in misunderstanding and a cost for the editors in trying to correct misunderstandings. I have yet to hear of much actual benefit for them, except that they enable round-trip mapping to GB 18030. :-( > > And then if a font designer chooses to provide a glyph and a mechanism > for displaying it when the corresponding ideographic description > sequence is encountered, whether in Han or in cuneiform, presumably that > is permitted. Ken, you wrote > > >There is no > >requirement for a conformant Unicode renderer to actually attempt > >a rendering of the Han character so described. > > > - but your wording implies that a renderer may do so if it wishes, > although with current technology more or less the only way it can do so > is to substitute a predefined glyph. Yes, it may do so. Just as a cuneiform renderer could take a description such as "SIGN1 TIMES SIGN2" and map it to a predefined glyph for the otherwise unencoded sign. Such behavior is simply a higher-level protocol defined on top of the text content. --Ken
