From xemacs-m  Sun Dec  1 04:39:11 1996
Received: from venus.Sun.COM (venus.Sun.COM [192.9.25.5]) by xemacs.cs.uiuc.edu (8.8.3/8.8.3) with SMTP id EAA24499 for <xemacs-beta@xemacs.org>; Sun, 1 Dec 1996 04:39:11 -0600 (CST)
Received: from Eng.Sun.COM ([129.146.1.25]) by venus.Sun.COM (SMI-8.6/mail.byaddr) with SMTP id CAA09474; Sun, 1 Dec 1996 02:38:06 -0800
Received: from kindra.eng.sun.com by Eng.Sun.COM (SMI-8.6/SMI-5.3)
	id CAA07696; Sun, 1 Dec 1996 02:38:04 -0800
Received: from xemacs.eng.sun.com by kindra.eng.sun.com (SMI-8.6/SMI-SVR4)
	id CAA20651; Sun, 1 Dec 1996 02:38:03 -0800
Received: by xemacs.eng.sun.com (SMI-8.6/SMI-SVR4)
	id CAA18639; Sun, 1 Dec 1996 02:38:00 -0800
Date: Sun, 1 Dec 1996 02:38:00 -0800
Message-Id: <199612011038.CAA18639@xemacs.eng.sun.com>
From: Martin Buchholz <mrb@Eng.Sun.COM>
To: mule@etl.go.jp
Cc: morioka@jaist.ac.jp, xemacs-beta@xemacs.org,
        Teruhiko Kurosaka <Teruhiko.Kurosaka@Japan.Sun.COM>,
        Bob Brewin <brewin@Eng.Sun.COM>
Subject: Re: mule API of Emacs and XEmacs (Re: caesar-region)
In-Reply-To: <199610150225.LAA10223@etlken.etl.go.jp>
References: <199610140521.WAA23056@xemacs.eng.sun.com>
	<199610150225.LAA10223@etlken.etl.go.jp>
Reply-To: Martin Buchholz <mrb@Eng.Sun.COM>
Mime-Version: 1.0 (generated by tm-edit 7.94)
Content-Type: text/plain; charset=US-ASCII

>>>>> "Kenichi" == Kenichi Handa <handa@etlken.etl.go.jp> writes:

Martin>    Date: Sun, 13 Oct 1996 22:21:21 -0700
Martin>    From: Martin Buchholz <mrb@Eng.Sun.COM>
Martin>    ...
Martin>    I think I need an education on this.  My requirements come mostly from
Martin>    the Sun Japanese Scientific Center, and they were quite vocal about
Martin>    wanting to see JISX0201 for all Roman text, in particular because it
Martin>    is impossible to get a Solaris ISO8859-1 monospaced font that is
Martin>    exactly half the width of the user's JISX0208 font, and they REALLY

Kenichi> ??? Both Sun's OpenWindow and XC's X contains fonts 8x16 (iso8859-1),
Kenichi> 8x16rk (jisx0201), 12x24 (iso8859-1), 12x24rk (jisx0201).

The Unix world is trying to move to a CDE environment.  Suppose we're
running CDE, and the user does the documented thing and use the CDE
style manager to select a font size (by `tee-shirt' size).

Then the default font for the emacs text area is set to something like:
-dt-interface user-medium-r-normal-s*-*-*-*-*-*-*-*-*

The fonts matching this are:

(mrb@xemacs) ~ $ xlsfonts -fn '-dt-interface user-medium-r-normal-s*'
-dt-interface user-medium-r-normal-s sans-12-120-72-72-m-70-iso8859-1
-dt-interface user-medium-r-normal-s-14-120-75-75-m-120-jisx0208.1983-0
-dt-interface user-medium-r-normal-s-14-120-75-75-m-60-jisx0201.1976-0


Notice how the jisx0208 font is exactly twice as wide as the the
jisx0201 font, but this is not true for the iso8859-1 font.

It will be tough to find an appropriate iso8859-1 font exactly half
the width of the jisx0208 font above.  Even if we do, we will have to
make a system-specific choice of font.  On Solaris, at least, the best
result for the user seems to be obtained from using the jisx0201 font
to display ASCII characters, assuming that one cares about preserving
the 2-1 display ratio.  Note that this makes column-dependent
functions work better visually as well, as long as Emacs estimates
Kanji as being twice the width of ASCII, as is traditional.

Martin>    If you have a buffer containing Japanese-Roman characters, and you
Martin>    type 'A', do you get an ASCII 'A' in the buffer, or a Japanese-Roman
Martin>    'A' ?  If an ASCII 'A', won't this look really weird?  To have a
Martin>    keypress insert different characters in different buffers would also
Martin>    be really weird.

Kenichi> I think all `A' in ISO 646 variants should have the same character
Kenichi> code in Emacs.  So, as I wrote in my previous mail, I'm planning to
Kenichi> map jisx0201-roman characters to ASCII except for backslash and tilda
Kenichi> which are recognized as jisx0201-roman characters and displayed by
Kenichi> jisx0201 font.  When one hit `\' on keyboard, backslash should be
Kenichi> inserted by default.  If one want to enter `yen' by hitting `\', we
Kenichi> can provide such a minor mode.

I agree that a typing backslash should insert backslash, but this is
hard to reconcile with using a jisx0201 font to display ASCII.
Interestingly enough, the Solaris jisx0201 fonts display a tilde as a
tilde, not as an overbar.  So it seems there is only one character to
worry about on this platform.  I wonder what the considerations are on
other systems running CDE?

Martin

Shell transcript follows:

(mrb@xemacs) ~ $ xrdb -query | grep 'interface user'
*DtEditor*textFontList:	-dt-interface user-medium-r-normal-s*-*-*-*-*-*-*-*-*:
*Font:	-dt-interface user-medium-r-normal-s*-*-*-*-*-*-*-*-*
*FontSet:	-dt-interface user-medium-r-normal-s*-*-*-*-*-*-*-*-*
*XmText*FontList:	-dt-interface user-medium-r-normal-s*-*-*-*-*-*-*-*-*:
*XmTextField*FontList:	-dt-interface user-medium-r-normal-s*-*-*-*-*-*-*-*-*:
*textFontList:	-dt-interface user-medium-r-normal-s*-*-*-*-*-*-*-*-*:
*userFont:	-dt-interface user-medium-r-normal-s*-*-*-*-*-*-*-*-*:
(mrb@xemacs) ~ $ xlsfonts -fn '-dt-interface user-medium-r-normal-s*'
-dt-interface user-medium-r-normal-m sans-14-140-72-72-m-90-iso8859-1
-dt-interface user-medium-r-normal-m-16-140-75-75-m-140-jisx0208.1983-0
-dt-interface user-medium-r-normal-m-16-140-75-75-m-70-jisx0201.1976-0
(mrb@xemacs) ~ $ xlsfonts -fn '*medium-r*-16-140-75-75-m-*'
-dt-gothic-medium-r-normal--16-140-75-75-m-140-jisx0208.1983-0
-dt-gothic-medium-r-normal--16-140-75-75-m-70-jisx0201.1976-0
-dt-interface user-medium-r-normal-m-16-140-75-75-m-140-jisx0208.1983-0
-dt-interface user-medium-r-normal-m-16-140-75-75-m-70-jisx0201.1976-0

