Subject: Re: multi-lingual support under Unix
From: hashao (email@example.com)
Date: Sat Dec 30 2000 - 11:09:37 CST
On Saturday, December 30, 2000, Mike Nordell wrote:
MN> Tomas Frydrych wrote:
>> What we will need is some kind of keyboard switching mechanism
>> that would allow to translate keyboard input into appropriate
>> unicode values based on a language/keyboard the user chooses.
MN> This is the responsibility of the OS/user-environment as I see it. It is not
MN> something every application should implement by itself (just think of the
pango (which will be the base of gtk1.4) will support switching input
method on the fly. So for X we can hope that gtk will handle it for
MN> As for implementation, I think it should only be allowed at the paragraph
MN> level. But as for spell-checking, I also believe we should have a flag (only
MN> dynamic/run-time) telling "the user don't want this word spell-checked".
As Karl Ove Hufthammer said in another mail, language tag should not
be limited to paragraph. It should be tagged where the user want it.
For spell-check thing, since CJK currently don't have an good
spelling check scheme, spelling check is disable for CJK text in
current abiword. Of course it is not based on language tag but some
rough latin - non-latin test routine. So we have the basic structure in place.
MN> For the runtime behaviour re. formatting, what do we really need?
MN> 1. A list of fonts (containing *all* the different attributes all fonts used
MN> in the doc have). This is referred to by its runtime-ID.
MN> We have a document containing (pseudo-doc)
MN> <Courier;10pt;normal>Hi all, I'm a <bold>bold</bold> word</courier>
MN> This would create two entries in the fonts container:
MN> #1. Courier, 10pt, normal
MN> #2. Courier, 10pt, bold
MN> or we could even implement it in a more efficient fashion, let #2 become
MN> #2. <inherits font #1>, overrides it with "bold".
I think this is how Word did it. inherit thingy.
-- Best regards, hashao mailto:firstname.lastname@example.org
This archive was generated by hypermail 2b25 : Sat Dec 30 2000 - 11:08:08 CST