Re: css3-lists: Coptic and Ionic Greek

On Mon, 16 Dec 2002, Christos Cheretakis wrote:
> 
> Using phrasing from the description of the other systems, here you 
> go. The steps do not follow my previous implementation. I'm trying to be 
> as close to the WD's phrasing. [...]

Wow thanks, that kicks ass! Excellent.


> 2.1 If there is no mapping for current digit, fallback to decimal
> 
> Please note that steps 2.1 is only valid for modern greek.

Because the ancient greek systems have no missing digits, right?

Seems odd that the newer system would be technically inferior...

This is the only system that is not defined for such an odd range of
numbers. Are you really sure that modern greek numbers are only defined in
the range:

   1..89, 100..189, 200..289, 300..389, 400..489, 500..589, 600..689,
   700..789, 800..889, 1000..1089, 1100..1189, 1200..1289, 1300..1389,
   1400..1489, 1500..1589, 1600..1689, 1700..1789, 1800..1889, [...] 
   889700..889789, 889800..889889.

...?

That seems incredibly inconvenient.


> 2.5 For readability, include space between groups.
>
> Also note that step 2.5 is my addition. Modern greek uses sigma+tau to 
> represent 6, but sigma and tau are also used to represent 200 and 300 
> respectively.

By "your addition" do you mean that normal modern greek doesn't include a
space and is ambiguous? Should the ancient greek forms have spaces?


> Space does not mean U+0020, as the numbers are written as 
> single words. It just means ...space.

How about U+00A0 non-breaking space? Or U+2007 figure space?


> ancient-greek
> lower-modern-greek
> upper-modern-greek

Are these your preferred names?

How about ancient-greek, lower-greek, upper-greek, lower-greek-alpha,
upper-greek-alpha?

Or, greek, lower-modern-greek, upper-modern-greek, lower-greek,
upper-greek?

How about 'greek' as a shorter name for the 'lower-modern-greek' variant,
so that the names are: ancient-greek, lower-modern-greek and greek,
upper-modern-greek, lower-greek, upper-greek?

What do other people think?


> 3. If number contains a thousands group, then prepend lower numeral sign.
> 4. Always append numeral sign.
> 
> numeral-sign is U+0374
> lower-numeral-sign is U+0375

Is "numeral-sign" a (sometimes absent) suffix, or is it part of the
number? If it is part of the number, what would be the typical suffix? (In
decimal, the suffix is '.' because lists go '1.', '2.', '3.', ...)


> And some examples, with some, well, well-known numbers. The first 
> column contains the names of the digits, the second their unicode 
> position and the third an iso-8859-7 representation

Cool, thanks. (FWIW, all i need are the codepoints. The rest means very
little to me. :-) I use the codepoints to include examples in the spec.)


> the number of the beast would be:
> ancient-greek: U+03C8 U+03BE U+03DB U+0374

766? :-)


> 200300 would be:
> lower-modern-greek: U+0375 U+03C3 space U+03C4 U+0374
> upper-modern-greek: U+0375 U+03A3 space U+03A4 U+0374

...(where the space is needed to differentiate 200300 from 6000).

So would U+2007 be ok for that space? It seems like a good match for
me. What do long greek books do for their page numbers?


> Note: 6 & 200300 could be ambiguous in modern greek, without the space. 

How? Surely the U+0375 makes it unambiguous.


What would 891918 be?

Is there an example on the net of th 'modern' systems in use? Or anywhere
in fact? I'd love to see this in practice, especially how they cope with
the ambiguities and the 90 and 900s.

-- 
Ian Hickson                                      )\._.,--....,'``.    fL
"meow"                                          /,   _.. \   _\  ;`._ ,.
http://index.hixie.ch/                         `._.-(,_..'--(,_..'`-.;.'

Received on Monday, 16 December 2002 17:38:03 UTC