|
|
Seeing errors, or visual errors, express themselves by the misinterpretation of some characters. The capital "I" (i) is for many font types similar to the lower "l" (L). E.g. with the Arial font type they are equal. Analogously there is a visual problem with lapsed printer ink. When the "r" and "n" characters are too close to each other, they look like a single "m". And of course the same applies for the other way around.
Audible errors are the most difficult errors from which to acquire statistical data and analyze. This lies in the fact that a textual robot simply cannot hear.
Two current examples of audible mistakes are the following:
CONGSTAR the common eharing mistakes are: CONGSTER and KONGSTAR. This is not stricty a hearing mistake, however due to the fact one cannot "hear" the difference, the words sound alike.
On the contrary, we have YELLOSTROM with YELLOWSTROM, which are pronounced quite differently.
There are many more possible mistakes based on hearing such as with "borrowed words" or words taken from another language. Implementing all possibilities will take some time. We are working on them and will have then online by September.
Typing too fast means making mistakes. The fast fingers errors are the ones where characters are omitted. This way EBAY becomes EBY.
In addition, mixtures from "Fast Finger" and "Fat Finger" like for example EWEBAY emerge frequently in our database.
The start of Typostats with a detailed analysis regarding Fat Finger (German "Dicker Daumen" = "Thick Thumb") succeeded.
We e.g. combined letters, like E in ESPRIT with surrounding letters 3, 4, R, D, S, W and made an evaluation - e.g. EWSPRIT.