(To see this message in a much better format -- follow this link....)
http://androidcommunity.com/forums/show ... p?p=133454

I think someone should port Dasher to Android.

Dasher is free software, that takes an extremely unique approach to achieve information efficient text entry. It allows lots of different input methods, and is very adaptable, and customizable.

Here is a description from the Dasher Website.

Dasher is an information-efficient text-entry interface, driven by natural continuous pointing gestures. Dasher is a competitive text-entry system wherever a full-size keyboard cannot be used - for example,

* when operating a computer one-handed, by joystick, touchscreen, trackball, or mouse;
* when operating a computer with zero hands (i.e., by head-mouse or by eyetracker);
* on a palmtop computer;
* on a wearable computer.

...using a mouse[/touchpad/trackball], experienced users can write at
39 words per minute.

Quick video, just to show it work
http://www.youtube.com/watch?v=0d6yIquOKQ0

An hour long video, that goes into detail, as to how it works, all the
different ways it can be used.
http://video.google.com/videoplay?docid ... A&q=dasher

Dasher on an IPAQ
http://www.youtube.com/watch?v=4jxhJjkw ... re=related




Also I thought I would point out, that Dasher already has support for all sorts of languages, so if someone ported Dasher to Android, it would give it another unique option for text input in all sorts of languages.

Here is a link that shows the language files already available for Dasher.
http://www.inference.phy.cam.ac.uk/dash ... ABETS.html






The Dasher Website
http://www.inference.phy.cam.ac.uk/dasher/

Sounds interesting.

But I imagine it could be hard to write texts in a rather formal language (which is as we all know quite common among mobile phone owners) or to use abbreviations. Thus, wouldn't it be better if the system was simply relying on the frequency of the letters? For instance, we'd have more space assigned to vowels whereas consonants would be a lot smaller. Also, the whitespace character would be as big as vowels.

Anyway, how should Dasher be implemented for the G1? I don't think it'd make any sense in case you held it upright (not enough space to have the full Dasher keyboard on the right side), but rather if you held it horizontally. A solution might be having Dasher displayed vertically if held upright (=> the letters are on a horizontal line instead of a vertical one), so that you also aim downwards with your finger?

I might it a try as soon as I've got a bit more time to dig into that Android dev stuff.

codethief wrote:Sounds interesting.

But I imagine it could be hard to write texts in a rather formal language (which is as we all know quite common among mobile phone owners) or to use abbreviations.

This is a quick vid I made just now, to show how dasher makes abbreviations easy.
Better quality:
http://screencast.com/t/ffXiC9q2
Same video on youtube (lower quality)
http://www.youtube.com/watch?v=nZHB-FbSHJo


Dasher operates off of a language file. (Which is litterally just text in a file) You could make a custom language file that includes lots of common txt abbreviations. This would make those show up just as much as real words.

In addition, Dasher by default automatically adds new things you write to the language file. So any words you use a lot automatically get more space. Since people generally use such a small set of words in texting, it would quickly get to a point where you can enter whole words rather quickly.

codethief wrote:Sounds interesting.
Thus, wouldn't it be better if the system was simply relying on the frequency of the letters? For instance, we'd have more space assigned to vowels whereas consonants would be a lot smaller. Also, the whitespace character would be as big as vowels.

Here is another quick video to show why it's better that dasher uses contextual frequency instead of just which letters are more common overall.
http://screencast.com/t/Orlm08NT


As you can see in the video, after i type "ty" dasher knows that the next letter is almost definitely a "p" followed by an "e". I would not want it to show the vowels big right after I enter "ty".

However as you notice at the beggining, After I type just "L" -- it does indeed show the vowels all very large, because those are the most likley next letters, based on what I just typed.

codethief wrote:Anyway, how should Dasher be implemented for the G1? I don't think it'd make any sense in case you held it upright (not enough space to have the full Dasher keyboard on the right side), but rather if you held it horizontally. A solution might be having Dasher displayed vertically if held upright (=> the letters are on a horizontal line instead of a vertical one), so that you also aim downwards with your finger?

The way I see being the most convinient, is for it to display vertically, with a text box on top. Basically like it is in the Dasher on an IPAQ video I linked.

I think it would be cool to be able to use the touch screen, but your finger would get in the way a lot, so I feel that the roller ball would be the best method of input. Controlling it with the roller ball would be much like using a touchpad like he does in the demo.



---

I would suggest watching the whole hour long video. There is a lot of information in there to answer a lot of questions you might have.

Also, I totally suggest you download dasher and play with it for just a bit -- you will get a feel for it. Watching the video does not really give you a sense of how it feels to use dasher.

Under the file menu in dasher you can select "Import training text" so that words you use a lot will be more common, and the letters to create those words will get more shelf space, at the correct times. Or even better, if you poke around alittle, you can just find the training text file,(its just a text file) and replace it completely with whatever you want.

doors.txt;10;15

There have been found even what are called say, on 'multiple alleles', that is to say, two or more different 'versions' and 'readings' -in addition to the normal, non-mutated one -of the same place in the chromosome code; that means not only two, but three or more alternatives in that particular one 'locus', any two of which are to each other in the relation 'dominant-recessive' when they occur simultaneously in their corresponding loci of the two homologous chromosomes. ,