Alternatives For Keyboards

Are there AlternativesForKeyboards? At least theoretically? How could they look like?

---

Proposal 1: Mouse with smart completion:

Actually, that's an interesting question: can a visual mouse interface be made to type words fast with a mouse? For example, click on common words listed at the top and letters on the bottom, which leads to cascading pop-down lists that can cascade the entire word if needed. There may need to be bi-directional cascading to fit. I'm skeptical, but won't rule it out just yet. Example opening screen:

  and for      // there would be more, this is just a quicky sample
  the to is
  -------------
  A B C D E F ...

--top

Consider the OneFingerKeyboard.

Consider ChordedKeyboards.


Metrics

Possible Metrics to Consider:


Existing systems:


Discussion

It's hard to prove that a keyboard is better just because it might be faster in a test. That's like saying MySQL is faster than some databases so it is better. But that isn't a rigorous study at all. It is really hard to show ANY satisfactory, rigorous, objective evidence anywhere on this wiki, so demanding it is almost futile.

I don't see those as equivalent. Human productivity and machine productivity are two different animals. But, RDBMS speed tests do indeed play a part in decisions. It is one factor among many possible. The more factors weighed, the better (in general).

First, this is on the wrong page and out of context. Top claimed that we could just do a typing test to see if the keyboard was useful in the page ThereIsNoObjectiveEvidenceThatKeyboardsAreUseful. It is not objective evidence that keyboards are useful or better for humans by just doing the test Top suggested in ThereIsNoObjectiveEvidenceThatKeyboardsAreUseful. Similarly, mysql is not useful for humans just because of a mysql speed test.

We can do a test and find Assembly and Machine code is objectively better since it is faster statistically and empirically than python. Doing all the tests in the world to prove python objectively better for a certain task in computing would take so many man years that it is not worth it.

Top's demands for objective evidence are copouts because you know that no one can provide ample empirical and objective evidence (it would take years, which is not practical, and even after years and years of results, you'd still deny it). One would have to be Statistics Society of USA or the Statistics Society of Britain to provide these empirical results. Even statistics authorities of a country would not be ample evidence, since statistics can lie, statistics can be faulty, and many statistics do not cover the entire world. Demanding empirical and objective evidence is Top's copout so that he doesn't have to admit or accept something as useful, even if it is so obviously useful.

As soon as something is shown useful, Top simply says "provide a biz app example for reporting table data" since that's the domain he works in. As if his specific domain is more important than all other people's wide varieties of domains on this wiki, and as if everyone using a computer programming language is using it for reporting tabular data. As if whether something reports tabular data deems whether it is useful or not in all domains.

That's not an accurate assessment of my approach. I only demand objective evidence for superiority when somebody claims their HobbyHorse is objectively *net* better. If I slip up on this rule, then please complain about it where it happens instead of generalizations, because I don't recall ever violating it. As described in ThereIsNoObjectiveEvidenceThatKeyboardsAreUseful, a speed test is one of many possible objective metrics. I never claimed it was sufficient by itself. And, this section should move to ThereIsNoObjectiveEvidenceThatKeyboardsAreUseful (EditHint). I will not address any further issues about how or when I expect evidence here. Such bickering belongs in that topic, not here. --top


CategoryUserInterface, CategoryKeyboard

EditText of this page (last edited January 13, 2009) or FindPage with title or text search