Re: chairless hamza

From: Khaled Hosny (khaledhosny@eglug.org)
Date: Fri Jan 04 2008 - 11:55:04 CST

  • Next message: arno: "Re: chairless hamza"

    On Fri, Jan 04, 2008 at 07:43:38AM -0800, John Hudson wrote:
    > Dear Khaled,
    >
    >> I'd say Yes. Relaying exclusively on font technology doesn't ensure
    >> integrity of text, we are in a situation of to different Hamzas every
    >> one has different expected behaviour. If both are encoded as the same
    >> character, how I'd type the word وبٱلءاخرة and send it to some
    >> one and
    >> expect him to receive it as a Quranic Hamza without relaying on some
    >> sort of rich text markup?
    >
    > But you are always going to need to rely on the recipient having an
    > appropriate font to render the text, and Quranic text is always going to
    > need more complex fonts than modern Arabic, just as, for example, Biblical
    > Greek and Hebrew texts need more complex fonts than moderhn Greek and
    > Hebrew. So one way or another you either need to specify the font in markup
    > or instruct the recipient what font to use.
    >
    > I am also concerned about integrity of text, but coming at the issue from
    > the other side: I'm concerned that the same letter not be encoded as
    > different characters based on its context, because that undermines the
    > integrity of the character/glyph model. Tom Milo has some good examples of
    > interaction of hamza with prefixes, and I don't think it makes sense to
    > change the spelling of a word at the character encoding level to obtain a
    > particular appearance. So I was responding mainly to Arno's suggestions re.
    > using different hamza characters in different contexts.
    >
    > If I understand your idea correctly, you would prefer to encode a new
    > 'Quranic hamza' character. This is a better idea.

    Exactly, I'm proposing a "Quranic Hamza" to be used for chairless Hamza
    in Quranic text, not different Hamzas according to context.

    >>> I consider the variance in usage a matter of display, i.e. I think it
    >>> would be a big mistake to use a different character for the hamza to
    >>> achieve a different appearance or shaping behaviour, as this reduces ease
    >>> of text comparison, spellchecking, sorting, etc.
    >
    >> I think Unicode actually did this before, in the case of Farsi Yeh for
    >> example.
    >
    > There are a number of similar instances, and many of them are, in
    > retrospect, bad ideas, introducing character level distinctions between
    > what are traditionally script style variations, e.g. the distinction in
    > form of isolated and final k in naskh and nasta'liq, which is now enshrined
    > as a character distinction for individual languages. But this last point is
    > an important one: these character distinction exist to address the
    > perceived needs of different languages and user communities, *not* to
    > handle different shaping of the same letter within a single language.

    I tend to think in Quran a special case here. Though it is definitely an
    Arabic text, it is traditionally treated differently, having its own
    writing rules based on the special need to keep it as close to its
    original (early) written form as possible. So, I think it is safe to
    consider Quranic Arabic a special variant of Arabic that might have
    writing rules that contradict with MSA, and those needs should be
    addressed by Unicode.

    >>> I suspect that inertia will decide this issue: it is possible to resolve
    >>> the display of transparent hamza at the font level on top of the existing
    >>> definition of U+0621, so I doubt if the UTC will see much impetus to make
    >>> changes that might cause problems for existing implementations.
    >
    >> For that reason, I think another Hamza is a necessity.
    >
    > A separate 'Quranic hamza' character is the best solution *if* a layout
    > solution is not considered sufficient. It has, of course, security issues
    > since, in a non-joining context it will be confuseable with the U+0621. The
    > UTC are very concerned about such issues, so any formal proposal would need
    > to address these.

    That is a good point that we need to think more about, I wonder how,
    say Cyrillic alternatives to Latin characters, are handled.

    >> As I relaying on font technology alone, doesn't guaranty any standard
    >> behaviour, more ever one can't ask font vendors to provide fonts that
    >> provide a, currently, non standard behaviour.

    This was supposed to be "As I said above, relying on..." :)

    > You can always ask font vendors :)
    >
    > Regards, John
    >
    >
    > PS. I would prefer to engage in these discussions on the public Unicode
    > list, rather than in private email, in the interests of open dialogue on
    > these important topics.

    Sorry, this was a mistake, I didn't intent to make it private.

    -- 
     Khaled Hosny
    
    




    This archive was generated by hypermail 2.1.5 : Fri Jan 04 2008 - 11:58:53 CST