|
2014-07-07
, 16:39
|
|
Posts: 568 |
Thanked: 969 times |
Joined on Dec 2009
@ Toronto
|
#62
|
The Following 4 Users Say Thank You to shallimus For This Useful Post: | ||
|
2014-07-07
, 16:52
|
Community Council |
Posts: 4,920 |
Thanked: 12,867 times |
Joined on May 2012
@ Southerrn Finland
|
#63
|
Changing the subject slightly: I wonder who answered "I am happy with the current vkbd" in the poll? Talk about low standards!
|
2014-07-07
, 17:28
|
|
Posts: 568 |
Thanked: 969 times |
Joined on Dec 2009
@ Toronto
|
#64
|
I tried to use it on N9 and it was such a POC it is unimaginable, I could not write a simple message without incredible amount of wrongly guessed words!!
So for me, swype is of no use.
The Following User Says Thank You to shallimus For This Useful Post: | ||
|
2014-07-07
, 18:59
|
Community Council |
Posts: 4,920 |
Thanked: 12,867 times |
Joined on May 2012
@ Southerrn Finland
|
#65
|
Hmm. Was this in English, or another language? I've only used Swype on N9 in English and I found the experience mostly very pleasing. One or two words consistently guessed wrong, but other than that it was (for me at least) a fast and accurate way to input text: both prose and more demanding stuff like passwords.
...and even if you don't like Swype, I don't see how hunt-and-peck is a good alternative, but then it's all about the user experience and that's very subjective
|
2014-07-07
, 19:08
|
Posts: 2,076 |
Thanked: 3,268 times |
Joined on Feb 2011
|
#66
|
Tried it with english, finnish and bash.
English was passable, finnish horrible and shell script writing impossible.
|
2014-07-08
, 03:38
|
Posts: 958 |
Thanked: 483 times |
Joined on May 2010
|
#67
|
@droll Levenstein distance is good for spelling mistakes if you can get individual keystrokes. In case of a "swype"-like keyboard it does not work (IMO) because it's not possible to isolate individual strokes.
The current algorithm is based on geometric properties of the curve drawn. It walks through the tree of all possible words (with scoring & heuristics applied very soon to trim the candidates tree) and try to assign a score to every possible word based on properties showing user "intention" (distance to key, speed, sharp turns ...).
|
2014-07-09
, 09:22
|
Posts: 951 |
Thanked: 2,344 times |
Joined on Jan 2012
@ UK
|
#68
|
|
2014-07-09
, 10:21
|
|
Posts: 1,055 |
Thanked: 4,107 times |
Joined on Oct 2009
@ Norway
|
#69
|
For now i have not got enough information to make a basic risk analysis, so the release is now on hold.
The Following 16 Users Say Thank You to w00t For This Useful Post: | ||
Bundyo, eber42, jalyst, Jordi, juiceme, Makeclick, mariusmssj, mattaustin, minimos, nodevel, P@t, rcolistete, shallimus, vitaminj, Watchmaker, zlatko |
|
2014-07-09
, 11:34
|
Guest |
Posts: n/a |
Thanked: 0 times |
Joined on
|
#70
|
Still, I don't see how we are any worse off than we were before mariusmssj got in touch with Swype.
Tags |
bettertxtentry, huntnpeck sucks, okboard, sailfish, swype |
Thread Tools | |
|
As for the second point ("making money" aka Lucrativité) it depends on your local law.
Obviously Swype will say no, and now you may have brought their attention.