TerseNet seems like it has some good ideas, and I had some similar ideas, too. I would do some things differently though, such as not loading even one small ima

TerseNet seems like it has some good ideas, and I had some similar ideas, too. I would do some things differently though, such as not loading even one small image automatically (everything must be loaded explicitly), not having such a small limit for text files (I think that many documents will be much longer), not using Unicode, and some other things differently too, but much of that is like what I would have wanted to make, and I could write some of my own ideas about it too, in future, maybe.