Bah, forget graffiti! It's too difficult; the computer should be able
to help you with it, but it doesn't. It essentially says "do this
unnecessary work". Using graffiti is like using a help file. (And I
have a pretty long rant on those somewhere).<br>
<br>
If a really clever person sat down and drew lots of pictures, he might
come up with a good messag-ease-inspired input method where fingers are
moved in patterns similar to real letters.<br>
<br>
Letters start being drawn from different places (or at least in
different directions), so instead of graffiti, it could be like
connect-the-dots. Start your letter in one point, a bunch of dots
appear that you can drag your finger to, with labels for what letters
those dots would lead to. For example, one dot may be labelled "I",
while another could be labelled "L U". You could drag to the "I" dot
and lift your finger right there to get the letter "I" (or keep
dragging if more dots exist for that letter, though there wouldn't be
much point). If you drag to the "L U" dot, new dots appear out of that
one for the letters "L" and "U"; drag your finger to the one you want,
as with "I".<br><br>
Graphically, there would be no static keyboard graphic; the keyboard
would not obscure the screen but rather expand (smoothly) when in use.
This would save screen space /and/ look cool.<br>
<br>
All the different letters could be tied into a nice little database
which would contain the paths for different letters. Still no ability
to lift one's finger, but since the input would be completely guided,
that would not hurt like it does with graffiti!<br>
To easily look up letters, paths in the database would have to
intersect, consistently, at specific points. I guess a grid (where the
size of each box is recorded somewhere) would work there. Thus, each letter in the database would have a whole bunch of grid points stored in the order they must be touched.<br>
A program could generate the point that one must be moving from to
reach a certain grid point for a letter. Otherwise, completely
unrelated letters would pop up from points that happen to intersect
with other points! (For example, I on point 0,-3 would require contact
last with point 0,3).<br>
Some kind of tree would have to exist, generated by a magical program
(once), using that direction information, saying what letters can be
led to from other letters. That would be entirely to speed things along
in the case of really big alphabets, so the computer doesn't have to
look through every single letter in the database. For example, "I" at 0,-3 leads to "L" and is the end of the path for "I". Point 0,3 leads to a lot of letters; "I", "L", "J", "U", etc. (Sorry, I seem to have changed the story from that first example :p As I said, we would need someone clever to sit down and figure out the connections).
<br><br>Bye,<br>-Dylan McCall<br><br><br><div><span class="gmail_quote">On 7/30/07, <b class="gmail_sendername">David Schlesinger</b> <<a href="mailto:David.Schlesinger@access-company.com">David.Schlesinger@access-company.com
</a>> wrote:</span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">>So by using fingers instead of a stylus we're not talking about the
<br>>same use case anyway.<br><br>That's certainly not clear to me.<br><br>>Are you policing this project for violations?<br><br>Not at all, that's a silly idea; as I've said, I simply _am_ obligated<br>
to point out when a trademark held by my employer is being misused, and<br>that's simply a condition of having the trademark.<br><br>My credentials aren't at issue: I've got a long-standing involvement in<br>the community, I've presented at both the Ottawa Linux Symposium and at
<br>GUADEC (of which ACCESS was a sponsor) this year, I'm on the GNOME<br>Foundation's Advisory Board and I'm the chair of the Linux Foundation's<br>Mobile Working Group and vice-chair of the Linux Phone Standards Forum's
<br>Architectural Working Group. I've been on this list pretty much since it<br>existed, and as Sean can confirm, I've had a long-term interest in the<br>project--he and I have chatted on numerous occasions, and while I'm
<br>still waiting for _my_ unit, I can hang on a while longer...<br><br>Let me add that there was = interest expressed by Mickey Lauer at GUADEC<br>week before last in using some of the "Hiker Project" components--which
<br>we have made available under an open source license--in OpenMoko,<br>something I'd certainly encourage.<br><br>Let's be clear: nobody's "violated" anything. I felt it was necessary to<br>give a heads-up that heading too far in the direction you seem to be
<br>trying to stake out (i.e. "Graffiti _shouldn't_ be a trademark" or<br>"We're using the term in some slightly nuanced way and claiming it's a<br>'different use case'") could be potentially troublesome, if the upshot
<br>of it were to wind up amounting to the creation of a "test case" of<br>either of those notions. If, say, I _didn't_ point this out and, based<br>on the discussion, someone started a handwriting-related project called
<br>"moko-graffiti" or something, that _would_ be troublesome. Just sayin'.<br><br>I think I've pretty much made the issues clear, and there doesn't seem<br>to be anything useful being added here. If anyone wants to delve further
<br>into this, they're welcome to email me off-list.<br><br><br>_______________________________________________<br>OpenMoko community mailing list<br><a href="mailto:community@lists.openmoko.org">community@lists.openmoko.org
</a><br><a href="http://lists.openmoko.org/mailman/listinfo/community">http://lists.openmoko.org/mailman/listinfo/community</a><br></blockquote></div><br>