Re: Thai/Devanagari in GNOME Terminal

[Chookij Vanatham <chookij vanatham eng sun com>]

Hi Robert,

] 
] On Mon, 23 Oct 2000, Chookij Vanatham wrote:
] 
] >    It tells me that the logical sequences of Thai characters in my Thai
] >    document files are different. The seq (1) is correct sequence and
] >    the seq (2) isn't correct. I know that something is not correct in
] >    my document.
] 
] Ah. Ok.
] 
] Well, my user tells me that input methods should prevent (2).
] That sounds good to me.
Again, what's if those sequences are in the text file already,
I still don't want to see them being displayed the same.

Again, input methods should not be confused with output methods.
Input methods can prevent these but there is one mode in that
input methods not to check any invalid sequence and I use it
to input invalid sequence to teach Thai kid that this's not
the right way to spell it. Hopefully, that Thai user person
wouldn't tell me not to input it...Right ? too much to control users.
Users come first.... :)


] 
] >    Also, if I want to sort these document, I would know that these 2
] >    sequences are not the same. The order is different. 
] 
] A decent collation algorithm will sort them equivilantly.
Does it mean that the sequence (2) which is not spelling correctly in Thai
is going to consider that it's equivalent to the seq (1) which is the correct
one ? 
Hmm!!! I think it's getting confused more, isn't it ?

] 
] However, I don't believe it's useful enough to break wcwidth(), as there
] are other ways of visually indicating errors than use of varying-width
] characters.
Well. Just to make sure that I'm not suggesting to break whatever we have,
particularly for wcwidth() to accommodate this for Thai.

I'm very open to see other ways too as long as, users's point of view
are ok. So do I.


] 
] > This idea isn't from me. It's from Thai Wtt2.0 standard specification
] > which was sponsored by NECTEC long time ago. It's done by a lot of
] > professional/experience persons both from technical engineers and
] > linguistic persons.
] 
] Which just goes to show, standards, no matter who designs them, can be
] broken. :)
That's very true. I agree, anything can be changed.
Like I said before, any methods to help clearify this, it's really ok for me. :)


] 
] I'm glad that this isn't anything actually required by valid Thai data.
Let me think more if I can find one... :)

Chookij V.