> > My keyboard, both at home and at work has the behaviour that, when I
> > hold down left-shift and press t and r keys simultaneously (or nearly
> > simultaneously), neither (or only one) of T and R appears, but not both.
> > Right shift has no problems. Other pairs of letters seem to work fine
> > with left-shift, but some don't: u and y, and p and [
> my sony picturebook shows this with
> P+{, !+@, #+$, ^+&, (+), P+:, !+Q, #+E, %+R, %+T, &+Y, *+U, *+I,
> (+O, a_+P, ++} and that's only testing keys that are next to each other,
> when i found that !+A shows this as well, i gave up.
This must have to do with the matrix-like wiring of the keys. There's
approximately an 8*16 matrix, with the keys in the crosspoints. The
keyboard controller scans a row or column at a time for pressed keys
(that's where the PC's scan code thing comes from - the keyboard was
effectively sending the key coordinates). As complication you have to
make sure that contact bounce is eliminated, and that pressing multiple
keys simultaneously is an error and no key code is generated, unless
it's a shift-type key (shift, ctrl, alt etc). Remember a long time ago
when keyboards were advertised with 2-key rollover and n-key lockout?
Back in those days I made my own keyboard (mega luxury version,
features guaranteed not commercially available), and implementing that
wasn't so easy. Getting it wrong produces the kind of effect you
describe. Location of the shifting keys in the matrix can have a
significant effect on the complexity of the programming required (if I
remember correctly).
Volker
--
Volker Kuhlmann is possibly list0570 with the domain in header
http://volker.dnsalias.net/ Please do not CC list postings to me.
