At 08:50 AM 10/11/02 -0700, Doug Ewell wrote:
What is the correct IBM GCGID value for U+03B8 GREEK SMALL LETTER THETA? Is it GT610000 or GT610002?The Unicode 1.1 lists (UNICHP6B.TXT and UNICHP6C.TXT) are inconsistent in this regard. Some entries, even within the same file, show GT610000 while others show GT610002. The tables printed in Unicode 1.0 book are the same. IBM has a Web page containing many PDF charts of code pages, and they have the same problem: some show one GCGID for U+03B8, others show the other one.
Wouldn't you be able to tell by the shape associated with the GCGID?
They were swapped in Unicode 3.0 / second edition ISO 10646 to make the alphabetic sequence match the usage in typical Greek Text using an ordinary serifed typestyle, reserving the other code point (U+03D1) for the symbol as it shows in mathematical usage. (Some text fonts will use a form matching 03D1 for 03B8, but that's OK - those fonts are fully usable for text, just not for math).I suppose this might have had something to do with confusion over U+03D1 GREEK THETA SYMBOL, but that character (a glyph variant of U+03B8) has been in Unicode since 1.0. Was there some dispute at that time over the preferred glyphs for U+03B8 and U+03D1? I remember that they were swapped in Unicode at one point. Was the inconsistency in GCGID a precursor to the decision to swap glyphs?
Historically, as far as I can tell, this relates to the fact that SC2 has documented its 8-bit character set standard with sans-serif type style(s) but 10646 and Unicode are using a serifed type style for the representative glyphs. Sans-serif fonts often contain the straight theta instead of the loopy one. As long as (small) character sets were only intended for text usage, any theta will do - Unicode and 10646 must be usable for both text *and* technical notation(s). That's where precise choice of representative glyph begins to matter. (The same is true for 0061 showing a hooked lower case a in distinction to the round lower case a for IPA. A font for non-IPA usage is free to use either form for 0061, but a font that needs to support or at least enable IPA usage is limited to the hooked form.).
Any ideas? (It's probably best not to ask why I am paying attention to GCGIDs in the first place.)
