Hi all,
Finally I got some time now to add further to what I had posted on 7th
Sept on this subject. Prior to that posting, Chintana Wilamuna had
queried to the effect that since SCIM + M17N combination had provided
Samanala key map for Sinhala from some time back, why the event of the
porting of AR's GTK Sinhala im module to SCIM is considered as something
new. Well, I wish to highlight some things that are distinct (and which
point to this new development achieved during the Foss week Code-Fest as
fully worthy of appreciations).
As those who are initiated with the use of SCIM would agree with me,
SCIM input method system gets implemented by two components: A Front end
and a Back end. The Front End is the SCIM with its configuration and
panel tools (including SKIM enhancements which is useful as a better
optimization for KDE environment; further SKIM is mandatory for Qt based
apps.) The back end of the system are the Input Method Engines (IMEs)
that one chooses to include. The official project site of SCIM
(including SKIM) is
http://www.scim-im.org and we can see at
http://www.scim-im.org/projects/imengines that the IMEs fall under the
following two categories:
1) Specific Language IMEngines
2) Multi-language IMEngines.
The M17N is one of the Multi-language IMEs.. It provides key map support
for number of languages including ta-itrans for Tamil and si-samanala
for Sinhala. (By the way, M17n [
www.m17n.org] project is conducted by
the AIST in Japan with funding by the Government of Japan!)
In contrast, a Specific Language IME is designed, as self explains, for
a single language. It can be seen in the site I mention above,
scim-sinhala (current version: scim-sinhala-0.0.0) which is the one that
was ported from AR's GTK im module, is a Specific Language IME. There
are couple of points to note:
a). Each of the other Specific Language IMEs available so far, is for a
language in CJK group (CJK = Chinese, Japanese & Korean) - thus
scim-sinhala is the first non-CJK to be tried out in development now
with specific language ime approach.
b). If one wants to use a Specific Language IME with SCIM (say compiling
from tar balls and installing), then all that are needed to be fetched
are the tarball for SCIM and after the installation of that, the scim
tarball for the needed language pack. (In case of scim-sinhala now it
is:
http://prdownloads.sourceforge.jp/scim-imengine/16422/scim-sinhala-0.0.0.tar.gz).
In contrast, for use of SCIM with M17N back-end, one needs 3 additional:
m17n-db, m17n-lib and scim-m17n. Well, installation-wise these are not
any more complicated than scim and of course from m17n one gets key maps
for a much larger number of languages. Nevertheless if I (rather a
newbie in transition to somewhat-knowing-somethings-level) have to give
instructions to someone who is much more newbie than me, to install scim
key map(s) for just 01 or a few more languages, I would hope that each
of those needed key maps are available as the more direct Specific
Language IMEs; because it would be much easier to start with lesser
requirements and not having to bother to explain about m17n at all.
In addition to above it is also noteworthy that the key mapping
mechanisms of m17n-si-samanala and the AR's GTK-IM / scim sinhala are
not mutually duplicating. Samanala scheme belongs to the qwerty based
phonetic types in which the basis for each consonant is the
corresponding generic character having no mixing with any vowel - that
is consonant characters with virama ("al-lakuna" in sinhala) symbol. In
the Scim-Sinhala key map type of qwerty based phonetic scheme
implementations, the basis of a consonant is the corresponding consonant
character having the first vowel ('ah' sound) mixed in - that is, ka,
pa, sa ... etc.
~Sethu
K Sethu wrote:
> On 9/7/05, *Chintana Wilamuna* <chintanaw@???
> <mailto:chintanaw@???>> wrote:
>
> On 9/7/05, Anuradha Ratnaweera <gnu.slash.linux@???
> <mailto:gnu.slash.linux@???>> wrote:
>
>> We all just witnessed a historic moment here. Kazuki Ohta has
> ported
>
> Not to underestimate anyone's effort but I thought this is old news.
> I'm using SCIM with M17N input module (package scim-m17n) to insert
> Sinhala Unicode characters. Yes it doesn't render them properly. (is
> it because the lack of surrounding text support ?)
>
> As you can see on the bottom right hand corner the keyboard layout is
> based on the Samanala scheme.
>
>
> Hi, in SuSe 9.3 I find si-samanala under SCIM+M17N on Gnome
> rendering fine but not in earlier Mandrake 10.1 & FC3.. - I guess
> (I am rather a newbie to all these!) it is because in my SuSe 9.3 it
> is recent Pango 1.8 whereas in other earlier distros older versions of
> Pango are present.
>
> In case of Tamil, in distros with older Pango there were some
> rendering problems in OOo but with Pango 1.8 in SuSe 9.3, OOo-2 works
> well.
>
> While SCIM works alright in GNOME as well as with many GTK apps, in
> case of KDE enviroment / and apps needing Qt I haven't found success
> with SKIM implementation yet. For example with KOffice in SuSe 9.3
> even with SKIM on, I do not find a way to launch the panel to make
> tamil / sinhala input.
>
> Since I do not know many Sinhala characters - can someone tell me if
> the .mim file for si-samanala is complete - that is covering all
> utf-8 Sinhala set (you have to view the file in Pango 1.8 or later
> enabled environment, I guess)
>
> ~Sethu
_______________________________________________
advocacy mailing list
advocacy@???
https://secure.lug.lk/mailman/listinfo/advocacy
(text/plain)