W3C home > Mailing lists > Public > public-i18n-core@w3.org > July to September 2007

Re: List-style-type: armenian in CSS 2.1 and CSS3

From: Simon Montagu <smontagu@smontagu.org>
Date: Wed, 01 Aug 2007 16:11:52 +0300
Message-ID: <46B08698.6070708@smontagu.org>
To: Richard Ishida <ishida@w3.org>
Cc: www-style@w3.org, public-i18n-core@w3.org, kode@hotbox.ru

I've filed https://bugzilla.mozilla.org/show_bug.cgi?id=390440 against 
Firefox/Mozilla but I'd like to have some clarifications made about the 
correct usage before writing or accepting a patch.

http://en.wikipedia.org/wiki/Armenian_numerals, which quotes no sources, 
corresponds to the implementation in Firefox and Opera (upper-case 
characters and only Ւ for 7000).

http://links.jstor.org/sici?sici=1539-5588(193905)13%3A8%3C368%3AAONSOA%3E2.0.CO%3B2-G 
is an article from National Mathematics Magazine, Vol. 13, No. 8 (May, 
1939). I don't have access to download the full article, but the URI 
shows the first page, which includes a table showing lower-case 
characters and only ւ for 7000.

Furthermore, there are contradictions in 
http://www.w3.org/TR/css3-lists/ -- the prose description of the 
algorithm says:

"This is a simple additive system defined for the range 1 to 99999999. 
The digits are split into two groups of four (if there are less than 
eight digits, the least significant group is filled first). Within each 
group, appropriate digits are picked from the following list (at most 
one per column) and written in descending order by value (thousands 
first). Any characters in the most significant group are then combined 
with a circumflex accent, ◌̂ U+0302."

This implies that the circumflex has the effect of multiplying by 10000, 
but the following example uses the circumflex to multiply by 1000:

"Example 1: Decimal 7482951 in lower-armenian is ու̂ն̂ձ̂սջծա U+0578 U+0582 
U+0302 U+0576 U+0302 U+0571 U+0302 U+057D U+057B U+056E U+0561. "

If the example is correct, the system will only be defined up to 
9,999,999 and not 99,999,999. Digits from 1000 to 9000 would also have 
two possible representations: either ռ ս ... or ա̂ բ̂ ... and it isn't 
clear whether one should be preferred or either may be used.

Best regards,
Simon Montagu

Richard Ishida wrote:
> I have updated the i18n Activity tests for list-style-type: armenian
> and run the test on latest Windows versions of IE, Firefox, Opera and
> Safari.  I'd be happy if anyone can run the test on other browsers on
> other platforms and report back to me (preferably with screen shots).
> 
> 
> Three of the four browsers tested (Firefox, Opera and Safari)
> supported armenian rendering of list-style-type as specified as far
> as 9999, as specified in the CSS3 module, except that:
> 
> 1. Firefox and Opera produced only one of the two characters
> specified for 7000 2. the rendering for the value armenian in all
> cases uses upper-case Armenian characters rather than the lower-case
> currently specified by the CSS3 module.
> 
>> From 10,000 onwards the results vary, but none follow the CSS3
>> spec. Whether this is ultimately significant is dubious in my mind,
>> since it seems unadvisable that ordered lists will be over 9,999
>> items long.
> 
> Test: http://www.w3.org/International/tests/sec-list-style-type-2 
> Results:
> http://www.w3.org/International/tests/results/results-list-style-type-armenian
> 
> 
> Can we say that armenian list style type has been implemented for CSS
> 2.1?
> 
> Since the wording is vague in CSS 2.1 I would argue yes from the
> algorithmic point of view, since the numbering works perfectly as far
> as 6,999 - which is a pretty long list already.
> 
> Wrt the upper-casing, this is clearly not intended by the CSS3 spec,
> which groups the value 'armenian' with 'lower-armenian', and if
> upper-case is accepted as the default the CSS3 Lists module will need
> to change.  On the other hand, upper-case is already consistently
> implemented across at least three browser implementations, so perhaps
> we should accept that as a de facto standard.
> 
> I'd like to hear from representatives of the browsers on this list as
> to whether they intend to change their implementations.  And I'd like
> to hear from IE representatives whether they intend to implement
> armenian list style types soon, and if so what case they intend to
> use. I think answers to those questions will help us move forward
> with armenian.
> 
> 
> RI
> 
> ============ Richard Ishida Internationalization Lead W3C (World Wide
> Web Consortium)
> 
> http://www.w3.org/People/Ishida/ http://www.w3.org/International/ 
> http://people.w3.org/rishida/blog/ 
> http://www.flickr.com/photos/ishida/
> 
> 
> 
Received on Wednesday, 1 August 2007 13:12:28 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 1 October 2008 10:18:52 GMT