(To scim-devel: This is a bug report from a Debian user. If necessary,
I'll submit a bug in SourceForge bug tracker as well.)
The original report is:
--- 8< ---
(From Michael Stroucken:)
I would like to enter the phrase "ååå" (gong'an ju) by typing
gonganju, but smart pinyin gets confused after 'gong' and takes the g o
n and g as separate characters. It is possible to pick the correct
character for 'gong' but it still sees the o n and g as remaining
candidates for interpretation. A workaround is to type gong'anju, but it
interrupts my flow :)
--- >8 ---
The full Debian bug report is
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=291045
On Thu, Feb 03, 2005 at 07:40:44PM +0800, Michael Stroucken wrote:
> Ming Hua wrote:
>
> >
> >Thanks for reporting. As you've said, the apostrophe is necessary in
> >case of ambiguity. In your example it seems ``gong an ju'' is not
> >ambiguous since ``gon'' is not a valid syllable. However scim is giving
> >the correct response according to your setting.
> >
> >You are seeing scim taking g, o, n and g as separate characters because
> >you had setting ``accept incomplete pinyin'' on. This is desireable in
> >cases people want to only type the consonants of a multi-char word. For
> >example, they can just type ``hypy'' for ``æèæé''. I personally
> >don't like this feature, but many others think it indispensable. And it
> >is on by default in scim-chinese. In this case, you get the character
> >``å'' just from ``g'', so scim still consider o, n and g not
> >interpreted.
> >
> >So if you don't want this feature, you can turn it off in setup,
> >IMEngine -> Smart Pinyin -> tab Pinyin. If you set it this way, you are
> >not going to get the remaining o, n and g. However scim still can't do
> >advance match as there is not ambiguity, but I think that's another
> >issue.
> >
> >Are you satisfied with this solution?
> >
> >
> Hmm, not really. But I just wanted upstream to know about the issue. My
> recollection may be incorrect, but I think the pinyin standard says
> apostrophes are only required when there is ambiguity, but there is no
> ambiguity in the string "gonganju". Scim should always display possible
> complete pinyin solutions even if incomplete pinyin is turned on. After
> all, I can mix incomplete pinyin within a set of words and it doesn't
> get confused.
>
> Maybe have something like:-
> 1) Either this letter is the last of previous character or
> 2) the first of the next character or
> 3) is part of incomplete pinyin
> and offer choices in that order.
I am not sure there is a sane solution for this, since if you allow
incomplete pinyin, there would be ambiguity for most of character
combinations, for example, ``ba'' can be a character or two characters
with the first incomplete, i.e. ``b'' and ``a''. Then you need to give
complete pinyin higher priority, and complicate things a lot.
But sure, this is a reasonable wishlist. Forwarding upstream.
Regards,
Ming
2005.02.03
