Touchscreen filters in andy-tracking...

[Andy Green]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Somebody in the thread at some point said:
| On Thu, Jan 29, 2009 at 18:03, Andy Green <andy at openmoko.com> wrote:
|
|> I gave Nelson the patch and a pointer to a PDF from TI that explained
|> the trick and computation needed, but even if we got it working I'm not
|> sure how much we can exploit the extra info.
|>
|
| Maybe such area information may be enough to to do some kind of 2
| fingers zooming ?
| -> increasing/decreasing global pressure area... (and only relative
| info is needed, so non normalization across the screen is not a big
| problem)
|
| I guess it cannot tell between horizontal rectangle area and vertical
| ? (rotation of 2 fingers).
| But anyway it would be much more difficult to implement rotation this
| way, with enough precision.

The information needed can't be had (from the existing measurements
anyway)... even the "area" number is just a single scalar, so you don't
know the aspect ratio of what's pressing down.  Then you can't "see"
rotation of the two fingers or guess that it is two fingers because the
aspect of the contact is non-square.

I can't remember how it acts with two contact points either, I'm not
sure it acted in a way that's good for this usage... it either picked
one or averaged them.

Maybe there's more tricks possible with the way the films are charged
and measured, but I can't think of what, the area measurement is pretty
tricksy already.

- -Andy
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org

iEYEARECAAYFAkmB510ACgkQOjLpvpq7dMrSMACffKGZnQGRhmMNFukaWfoWyqBu
3UQAn3wu72PFztLRYJtjnWNGJ4wbxHqi
=b38j
-----END PGP SIGNATURE-----