Hi Lex,

> I don't understand the issue.  What do you mean by the two items about
> being unique and about keyboards providing what one would think they
> should?

For being unique, there are really three values for all of the "duplicate
meta keys", namely a "normal one" (shift, alt, ctrl) as well as left-key and
right-key variants. And I am simply not sure if the use of these codes is
consistent across all machines or not - at least my keyboard seemed to
generate "alt" when I pressed "left-alt".

> As a discussion point, have you considered how to rewrite
> ParagraphEditor under the "raw codes" approach?  Would it work out okay?
>  If PE has to use raw platform-specific codes in key events, then it
> would seem to need separate handling for each encoding of raw events
> that it might see.  It would also need a way to find out what encoding
> is being used -- should it simply consult the platform string name?

That'd be a reasonable way of doing it and the translation could actually
happen earlier - for example right in Sensor which could do the translation
for all of the "well-known" codes in Squeak. This would leave the ability
open to support some of the not-so-well known ones, including for example
"OEM keys" which can be completely idiosynchratic.

> The situation looks funny from a distance, because the XWindows encoding
> has served Linux/x86 well, and Linux/x86 uses the same keyboard that MS
> Windows does.  Of course, XWindows may be getting lower-level events
> than are available via MS Windows.

That's quite likely.

> By the way, is there a listing of the Windows key encodings anywhere?  I
> would love to take a gander at it some time.

Yes, here:
http://msdn.microsoft.com/library/default.asp?url=/library/en-us/winui/winui/windowsuserinterface/userinput/virtualkeycodes.asp

Cheers,
  - Andreas