I have a software that is page oriented instead of scrollbar oriented so i can easily count the words, but i'd like a way to filter outliers and some default value for the text language (that is known).

The goal is from the remaining text to calculate the remaining time. I'm not sure what is the best unit to use. WPM (words per minute) from here seems very fuzzy and human oriented. Besides i don't know how many "words" remain in the text. http://www.sfsu.edu/~testing/CalReadRate.htm

So i came up with this:

The user is reading the text. The total text size in characters is known.

His position in the text is known. So the remaining characters to read is also known.

If a language has a median word length of say 5 chars, then if i had a WPM speed for the user, i could calculate the remaining time. 3 things are needed for this:

1) A table of the median word length of the language.

2) A table of the median WPM of a median user per language.

3) Update the WPM to fit the user as data becomes available, filtering outliers.

However i can't find these tables. And i'm not sure how precise it is assuming median word length.