An essay writer application could allow children around the seventh grade to learn to write better essays by being able to select some sentences from a database of sentences for a given topic. Psychological effects are that the children will experience success in essay writing because of the higher quality of the work and that the children will pay more attention to the (already correct) text because they can view it as their own work.
The application can also keep track of sentences that are compatible or incompatible so that the essay writing application can already make a decision about the quality of thought that has gone into essay writing. Some sentences could also offer spelling mistakes or other mistakes so that the children would still sometimes be able to make mistakes when selecting sentences.
 Availability of sentences
Sentences could be made available like a technology tree in a construction and management simulation, making some groups of sentences only available after a certain quality of writing had been established or certain groups of sentences or sub-topics had been selected. The intended psychological effect is to motivate and to teach continual refinement of text, including the removal of obsolete sentences, and reflection about phrases the child can understand but might not have used (passive vocabulary vs. active vocabulary).
A random generator could present sentences from a pool of currently available sentences for a given paragraph or section of an essay. The pool would be presented in random order but could, at least in theory, be completely viewed by clicking a "more" button. A search function could allow to search for certain words in the pool of available sentences. The sentences should lack citations because that is the purpose of the Research terminal. The assessment process could also recommend new section titles on occasion, which could make entire sub-topics only available after the corresponding section title had been granted.
Interviews with virtual experts could allow to improve the quality of a sentence pool through questions directed to the expert of choice (who might respond in a virtual email or might hold a short lecture on the selected topic).
Templates could allow to turn some sentences into multiple-choice questions or cloze inside the essay.
Half sentence templates could allow to combine the wrong half sentences.
 Improvement markers
A user could be able to place a limited number of improvement markers, which would instruct the virtual coeditor or head editor to search for better phrases. The process should be artificially slow to give the pupil motivation and opportunity to think for himself.
 Research terminal
Pupils could have access to a research terminal (a virtual search engine with limited content) with links to relevant publications about the topic and related topics.
The research terminal can also require that certain words have to be added to a word list of available words for search expressions. That could mean that one article has to be read, a relevant word has to be added to the word list by clicking the word (it cannot be typed) and then the word becomes available and allows to find another article that has previously been unavailable. The word list is meant as the relevant vocabulary for the essay and can be attached to the final document (as a glossary or as an attachment).
The research terminal could also allow to mark sentences from publications for use as citations.
A database could initially be filled with sentences from selected Wikipedia articles, including sentences from similar articles or topics and variants of sentences which may not be appropriate for the essay.
Different modes could allow very low error rates or very high (natural) error rates. With very low error rates a child would primarily train the thought processes required for the top level design of the essay (or another exercise) while with natural error rates the child could still benefit from proposals but would train all the skills required to prevent spelling mistakes and grammatic mistakes, almost as if writing the text unaided.
Different freestyle modes could allow more sentences to be inserted or rewritten by the pupil without the need to select templates or sentences from the database.
 Math trainer
One could implement a math trainer where pupils had to pick transformations of an equation or aspects of transforming operations from a database of possible ideas (most of them wrong or out of context, of course), which would allow the math trainer application to make a very thorough analysis of the transformations that had been applied appropriately and correctly. Mathematics teachers and classroom assistants would be able to offload some of the more time consuming and confusing tasks of having to follow irregular reasoning on to the computer and could focus on teaching to improve the reasoning of their pupils.
The math trainer can also directly link to tutorials that train the aspects of operations that have been recognized as not understood. If the application can invite to make certain mistakes the mistakes have been categorized and can be described in individual tutorials. Teachers and assistants might consequently be able to focus on helping to understand the tutorials.
 See also
 External links
- Application monitoring (Assistant teacher program, Wikiversity)