Michael Everson raised a very interesting question, which caused me to sit and think about it for quite a while.
>At 08:16 +0100 2002-06-24, William Overington wrote: > >>U+E7C2 HOLLY LEAF (GREEN) SURROUNDED BY FIVE BERRIES (RED) > >As a "character", will this differ from HOLLY LEAF SURROUNDED BY FIVE >BERRIES in its semantics? If not, then you are using character coding >for a higher level protocol again. Well, after some thought as to whether it would differ in its semantics or whether it would not differ in its semantics, I realized that I had no intention that HOLLY LEAF SURROUNDED BY FIVE BERRIES would be defined as well. I suggested U+E7C2 HOLLY LEAF (GREEN) SURROUNDED BY FIVE BERRIES (RED) as a test item, my reason being that the examples of chromatic characters that had been discussed are rather special characters and I, and probably others, would not wish to go around using them for tests in a manner that would possibly be usage in an incorrect manner, so an ornament that could be used in a general manner seemed desirable. Also, I felt that a specific example and some specific colour suggestions might help get something going. I suppose that HOLLY LEAF (GREEN) SURROUNDED BY FIVE BERRIES (RED) would appear in monochrome if a chromatic font were used on a platform which did not support chromatic fonts. I find it interesting to compare and contrast HOLLY LEAF (GREEN) SURROUNDED BY FIVE BERRIES (RED) where colours are defined within the name of the symbol with present regular Unicode where colour is not defined within the name of the symbol. If a chromatic font of letters of the alphabet were produced, the letter A would still be U+0041. Yet some symbols such as the holly would probably need colour guidance, either in their names or in notes in the code charts. I do not purport to have the answer to this, if the technology of chromatic fonts can take off, then it is an issue that would need to be resolved by discussions. Certainly, this example works because two of the colours which I have suggested as default colours are used in the name of the symbol. If two specific colours were needed where they are not two of the default colours which I have suggested, then I am unsure, at present, which way to proceed. I suppose that it is possible that the encoding of the character in the font could carry a colour specification with it so as to override any default settings of the rendering system for that character, with a "strength of assertion" value as to whether the override is merely suggested or obligatory. The rendering system could then have a "strength of defaults" value, with the arrangement that the default colours of the rendering engine would be overridden by the colour specification of the character in the font if the strength of assertion is greater than the strength of defaults, for that character within that rendering system. The strength of assertion and strength of defaults value could be an integer, composed of, say, three bits, giving a number from 0 to 7. So maybe the marks from the manuscripts would need to have a strength of assertion of 7 as they are obligatory, whereas something like the holly symbol could have a strength of assertion of 4 and a two colour capital letter could have a strength of assertion of 2. If a rendering system had its strength of defaults value set at 3, then the manuscript marks and the holly would show in their specified colours, yet a capital letter would show in default colours. These are just examples, for perhaps a rendering system would have its strength of defaults value set at 0 for normal usage and only be increased if there were some problem with that. William Overington 25 June 2002