Linas Vepštas

The above comment points out that we have an API issue. The current API is driven by word indexes: all words have a fixed location in a sentence, which means...

>> if a disjunct has a connector starting with z, that connector is ignored during pruning -- its "invisible". As if it was never there. > If seems z connectors...

> Moreover, they may get inserted in places that make invalid sentences parsable. For example: *`I will it, if she goes first` I don't see what would prevent some phantom...

> I guess it would be a good idea to learn the API of NLTK and mimic it when applicable. Given that it seems to be immensely popular, I suppose...

> What is the purpose parsing with islands? It's a flag that dates back to the original code. It says, basically "I can't parse the whole thing, but here are...

> 1. Are the phantom words supposed to always fit in island boundaries? It sure seems like it, doesn't it? > 2. Would this method of finding phantom words is...

>> * How should it be drawn, in ascii-art? Conceptually, the zero-verb is kind of floating in outer space, not having any linear position in the sentence. It only has...

That's right. github automatically creates these incorrect tar files, and I don't know how to make it stop. It's a whizzy feature that someone who works for github thinks is...

It would be better if I could just disable toe github release mechanism entirely. That last thing I need is to have yet more complexity and make-work in the release...

I'll think about this tomorrow. My knee-jerk reaction is that this is a bug, probably due to laziness and maybe forgetfulness, when the sql dict was implemented. Yes, I suppose...