Martin Thoma

Workers are only a special type of user. ## Step 1 Get rid of `wm_workers`. Add the contents to `wm_users`? Open questions: - What to do with `API_key`? - Simply...

When I want to add [fish 🐟](http://www.fileformat.info/info/unicode/char/1f41f/index.htm) to [this](http://www.martin-thoma.de/write-math/view/?raw_data_id=329579) I get > Fatal error: Uncaught exception 'PDOException' with message 'SQLSTATE[23000]: Integrity constraint violation: 1062 Duplicate entry '' for key 'formula_name''...

A language model could vastly improve the classification performance. To create one, I will parse - the (english) Wikipedia for `...` - the following LaTeX documents for `align`-environment, `$...$`, `$$...$$`...

http://www.martin-thoma.de/write-math/symbol/?id=547 `\boxminus` in amssymb and mathabx

Currently, the gallery page is too slow, because of queries like ``` sql SELECT `wm_raw_draw_data`.`id`, `data` as `image`, `creation_date`, COUNT(`wm_partial_answer`.`recording_id`) as `answers`, COUNT(`wm_worker_answers`.`raw_data_id`) as `answers2` FROM `wm_raw_draw_data` LEFT OUTER JOIN...

- [`\textnumero`](http://www.martin-thoma.de/write-math/symbol/?id=500): 94 recordings, 3-4 strokes

- https://developers.google.com/identity - http://merbist.com/2012/04/04/building-and-implementing-a-single-sign-on-solution/ ## StackExchange StackExchange OAuth has the big advantage that I can access data from the tex.SE site. If a user has earned reputation there, he knows...

Given a symbol class, find candidates which are likely to be missclassified. Currently, this is done with DTW. However, the results of this method are not satisfying.