After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 136482 - Consider two new keyboards: Navigation & 10-key
Consider two new keyboards: Navigation & 10-key
Status: RESOLVED FIXED
Product: gok
Classification: Deprecated
Component: general
unspecified
Other All
: Normal enhancement
: ---
Assigned To: David Bolter
David Bolter
Depends on:
Blocks:
 
 
Reported: 2004-03-07 19:58 UTC by korn
Modified: 2004-12-22 21:47 UTC
See Also:
GNOME target: ---
GNOME version: Unversioned Enhancement


Attachments
xml.in file for numberpad keyboard, including Tab and Shift-Tab (untested) (4.00 KB, patch)
2004-03-07 23:40 UTC, bill.haneman
none Details | Review
more correct numberpad keyboard (remember to intltool-merge it) (4.00 KB, patch)
2004-03-08 00:07 UTC, bill.haneman
none Details | Review

Description korn 2004-03-07 19:58:48 UTC
Consider adding two new keyboards to GOK.

Navigation:  UI Grab & Activate together do a great job making a large
number of tasks very efficient.  Nontheless, there are places where UI Grab
fails, or isn't the right approach.  Consider adding a Navigation keyboard
to GOK which contains "Tab", "Shift-Tab" (or "Back-Tab"), the arrow keys,
"PgUp", "PgDn", "Home", "End", and "ESC".  Perhaps one or two other keys as
well.  

10-key entry: there are a number of situations in which a user is only
entering numbers.  Doing this with the existing Compose keyboard,
especially as a single-switch user in the traditional layout (with the
numbers at the far right) isn't very efficient.  For this situation, a
10-key keyboard (or really the full typical 17-key Numeric Keypad) would be
a much better choice.  Down the road it may be possible for an application
to indicate to GOK that a particular text-entry field is a "numbers only"
field, in which case GOK might switch to this keyboard automatically.
Comment 1 bill.haneman 2004-03-07 20:32:32 UTC
I'd call the "10key" keyboard a "numeric" keyboard.  Seems like a good
idea, and one that we've considered.

Of course it's easy to create such a keyboard with GOK's keyboard
editor (now), but having one built-in and/or accessible from the
"standard" installed GOK keyboard suite would be nice.

There are two ways to go here: 

1) hard-code one, similar to the hard-coded "qwerty" keyboard;
2) build one from the current XKB numberpad keyboard geometry "section".

#2 would be more elegant, but it's not clear how easy it is to
identify the numberpad section.  It would also require that the
current XKB keyboard definition has a numberpad section (most do).  
#2 would also ensure that the numberpad keyboard was properly
localized, since the exact placement of numberpad keys and symbols may
not be fully standardized.  #1 would be really easy to do of course.

I'd suggest combining these two keyboards into one keyboard, with a
"shift" or "numlock" virtual key which would allow the same GOK
keyboard to do navigation and number entry.  Again, this would be
pretty simple.
Comment 2 bill.haneman 2004-03-07 20:40:10 UTC
This would be a 2.7.X feature, probably, though perhaps we could
bundle a hard-coded numeric keyboard with 2.6.1, strictly as a sample
(i.e. not connecting it to the default GOK keyboard set).
Comment 3 bill.haneman 2004-03-07 23:40:50 UTC
Created attachment 25310 [details] [review]
xml.in file for numberpad keyboard, including Tab and Shift-Tab (untested)
Comment 4 bill.haneman 2004-03-07 23:53:23 UTC
the attached keyboard is pretty rough (several obvious bugs) but I
just hacked it up fast.  It conforms approx to ISO 9995-4, so there is
in fact a "standard" applicable for numeric keypads, suggesting that
we might not have to extract this from the XKB description.
Comment 5 bill.haneman 2004-03-07 23:57:51 UTC
BTW I got the shifted-vs-unshifted keycaps backwards in the
attachment.  I also notice that GOK's "qwerty" keyboards keycaps don't
update with the Shift modifier (though the output does properly shift
case).  

will file a bug on that.
Comment 6 bill.haneman 2004-03-08 00:07:05 UTC
Created attachment 25312 [details] [review]
more correct numberpad keyboard (remember to intltool-merge it)
Comment 7 bill.haneman 2004-04-21 22:45:15 UTC
Remaining question is when to branch to this keyboard; I suggest we provide it
as a branch target on a gok keyboard, but which one?

Should we add a key to the 'GOK' keyboard?  Or add one to the bottom row of
Compose (currently used only for Text Manipulation?  David?
Comment 8 korn 2004-04-21 22:50:24 UTC
My vote would be on the Text Manipulation row.  I haven't tried any of the
recent patches to support Chinese; how does the user indicate s/he wants to
input in a different language?  Is it only based on the user's locale?  I bring
this up here because putting another key on the Text Manipulation row for
switching text-entry languages might also make sense...
Comment 9 bill.haneman 2004-04-22 09:51:02 UTC
Just as any user would use the Keyboard Indicator applet, GOK users can use it
to change their keyboard definition as seen by the X server, etc.

For the "special" languages which really require input methods or something
beyond what an XKB keyboard definition can provide (primarily Chinese, and
Japanese for users who don't wish to use only Hiregana or Katakana), this is
governed by a combination of locale (which can be set on the fly in the shell),
and user-specified "compose" keyboard (see bug #136454).  At the moment,
user-specified compose keyboards must be changed in the Settings dialog, but of
course any user wishing to create a "my keyboards" keyboard may do so, and
introduce that into the GOK branching keyboards list in any way s/he likes.

In practice, a locale-specific custom branching keyboard probably will want a
branch key which takes it to the XKB-specified compose keyboard as well, so that
a user may use physical keyboard emulation as well as GOK-input-method character
insertion.
Comment 10 bill.haneman 2004-04-22 17:19:31 UTC
fixed in CVS