I has been a while since my last post here. Development of Detexify had stalled because I was supposed to write a scientific paper about detexify and after that was done I needed a break so I spent my time on other projects.
Since a few days though I am working on detexify again. At the moment the changes are internal but they will ultimately lead to a big new feature. I’ve been asked a lot for characters like ä, ö, å, Å, etc. and why they are missing completely from detexify. In LaTeX these characters with accents or diacritics are all created similarly. The problem with adding them to detexify is that I would have to add more than fifty symbols for each accent (a-z, A-…). This would totally pollute the search space. But there is another way to add diacritics to detexify. I will add a second canvas which will look slightly different with a gray area in the center to represent an arbitrary glyph. A user can draw the diacritic he is looking for onto the canvas and attach it to the gray area (see mockup below) to get a list of matching commands and see how to produce the accented letter in LaTeX just like it works with normal symbols in the current version of detexify.
The architectural changes for this feature are nearly done and I hope to be able to release it soon. This will also open up the possibility for a unicode recognizer. Someone already did some work in that area and here is the forked project: http://github.com/nomeata/detexify/.
