I'm still thinking about how to transition the Khan model to an English Class. I think a regular High School English class wouldn't be so hard. You could offer a lecture from a teacher that was familiar with a selected reading, and perhaps make a series of videos covering different sections of the reading, or different chapters of a book. For exercises you could assemble multiple choice style tests.
The ESL classroom could also use some effective software. I am working on 2 videos, one for simple sentences, and a second for a simple reading.
The simple reading will be a video of a short text, including dialog, and followed by a walkthrough of three questions: one easy, one medium, and one hard. Linked to the video will be 6-9 more questions, following the models shown in the video. The problem with this model is that the questions will be hard to verify. I am trying to use google.docs and studyblue.com to build and accompanying set of study cards and a quiz.
This is more of a big idea than an actual possibility for me to pull off alone. I am think of the ESL market around the world. So many people out there are eager to learn English. Therefore, a very simple topic might be very popular.
I thought about introducing a simple sentence, such as "The boy kicked the ball." With pictures I could define the words, and then I could have the learner practice writing their own sentence following the video. The difficulty would be checking the sentences.
My idea is to build a database. Language is very versatile, so when you ask a student to write a sentence, even if you restrict them to 10 words and ask them to use "boy" and "kicked", you still have many possiblities.
The boy kicked the ball.
The boy kicked the cat.
The boy kicked the dog.
The man kicked the boy.
If we could pilot this program somewhere, then initially we would have to build a list of correct answers. We could take the pilot students answers and enter then in a database, flagging the correct answers. This would be time consuming at first, but hopefully we would begin to see some similarities between correct answers, and the program would become automated. The date would also server the teacher well in helping make follow up videos on common mistakes. This could be a very powerful system, but it requires a lot of ground work to start. Once it gets going though, I feel like it can accelerate learning.
I am still plugging away. Before I was beginning each day with review exercises. Now I am going to let them pile up, while I chip away at the reamining 20 or so exercieses I have left. On the weekend I am going to take a run at the reviews, and see if I can't knock them all off at once.
The vocabulary.com wesbite is a bit testy sometimes. I have 2 other people working on the challenge now, but they are 2nd Language learners, so they are not looking to speed towards the final goal, rather, just improve their vocabulary steadily.
I have 435,000 or so points. I am almost halfway there. When the site goes down I love the fervor, and sometimes it doesn't come back for a while.
The other day I explored answering questions wrong in favor of gaining points more quickly. This is worth investigating, but even if this "game" is "broken", I think the integrity of the tool stands. It is modeled as a self-pace tool to start with, so if you want to learn, you will use it correctly.