Hi Team
As we discussed yesterday, yes, its good to have some request-driven development and
update on this issue, suggested by ankit. And I have one more suggestion. I think, we need
to define or make some common rule to validate "significant difference in conjunct
glyph", when dependent vowel makes 4th level of conjunction on top of a conjunct
character. Few ideas for the rules in my mind are :
1. Consider, if the 3rd level conjunct shape totally change with adding 4th level
dependent vowel.
2. Consider, if normal conjuncting point of any above-base or below-base character differs
consonent to consonent or conjunct
consonent. For example, in bengali, dependent vowel sign U when adds with all
consonent, places itself just below the base
character, but when it conjuncts with RA, it places itself on a side of base character.
Although these cases are rare.
3. Thirdly, as none of us are multilingual experts, so, we can discuss and collect some of
these exceptional data sets from
different language experts and extend our rule set accodingly.
Regards
Satya
Quality Engineer
Red Hat Inc.
PNQ Center, India.
DID : +91 20 400 57385
Cell : +91 9860 140303
Fax : +91 20 660 56040
***The wise man should restrain his senses like the crane and accomplish his purpose with
due knowledge of his place, time and ability***
----- Original Message -----
From: "Ankitkumar Rameshchandra Patel" <ankit(a)redhat.com>
To: utrrs-devel(a)lists.fedorahosted.org
Sent: Thursday, May 31, 2012 10:50:08 AM
Subject: Re: Need for some more references test cases in UTRRS
Hi,
I have some suggestion based on
https://bugzilla.redhat.com/show_bug.cgi?id=825081 for Kannada
script. I observed in Kannada script after applying vowels U, UU
significant change happens in overall rendering.
We need some guideline. Since at one hand we have limitation for not
adding x*x combinations and same time we have to add all those
combination creating confusion to developers.
Proposed guideline: Add <cons+haln+cons+matra> under condition when
applying <matra> to <cons+haln+cons> combination create
signification difference/variance than applying <matra> applied to
single <cons>.
Regards, Pravin Satpute
I agree with the issue mentioned by Pravin, and to me the solution looks
a bit different.
Build a process of adding a combination via requests received from
developers or whoever has doubt about glyphs and combination. If the
combination doesn't exist in our system, we will add them and make it
available for general usage.
How about this?
--
Regards,
Ankit Patel
http://www.ankit644.com/
_______________________________________________
utrrs-devel mailing list
utrrs-devel(a)lists.fedorahosted.org
https://fedorahosted.org/mailman/listinfo/utrrs-devel