2012/5/29 Doug Ewell <doug_at_ewellic.org>:
> Philippe Verdy <verdy underscore p at wanadoo dot fr> wrote:
> Did you read what I wrote? The *underlying architecture* of Windows key
> handling supports neither additional shift states nor multiple dead
> keys, both of which are required to support this standard. A new version
> of MSKLC on top of the existing architecture will not help.
If that was true, then Keyman would not even work on Windows !
There are several layers in the Windows keyboard architecture (between
the physical keycodes detected at the hardware level, their
translation into virtual keycodes by the low-level device driver,
their classification into system keys handled directly and not
converted within Windows and the rest of the conversion to characters
which is fully part of the keyboard driver, then for their conversion
to applciations using several interface layers for BIOS/DOS
compatiblity, or for onversion to legacy non-Unicode "ANSI"
applications) and all these layers allow this support.
MSKLC provide the support in the driver it generates for almost all
these levels : it does not just allows building the layout tables, it
also creates a complete driver that integrates this layout. MSKLC does
not just build a layout file (this does not work alone on Windows), as
simply as what you can find on keyboard layouts for X11 on
Linux/Unix'es.
Keyman is different because it installs a generic driver that can then
load multiple layout files. The drivers built by MSKLC are not
separatable between the code part and the layout part, even if you
can create a source file describing only the layout data (these files
are not installable on Windows, they have to be compiled into a
driver, and the driver installed like a regular device driver).
Received on Tue May 29 2012 - 14:33:34 CDT
This archive was generated by hypermail 2.2.0 : Tue May 29 2012 - 14:33:34 CDT