Linas Vepštas

Aaron: > I suggest that we try a principled, fore-thought-out version! Yes, no complaint there. I was assuming that this was a very narrow-issue PLN-specific bug report, not an anything-goes...

Aaron: > ... the current pipeline doing it too early or too late ... > ... its way past time to sketch out the semantic processing layers of the pre-AGI...

Please, thanks, I can't overstate how important this is. In a certain sense, its more important than the code.

true about "open-ended issues never close." But there are some semi-specific topics, and perhaps 2-3-4 people should read the docs, try to run the examples, pose questions as they come...

This sounds a lot like one of those random Ben-thoughts. I am not convinced we should be writing down this sort of stuff, and putting it into github issues. Its...

This is vague and difficult to understand. Its also not really a bug, but an enhancement and should be tagged as such

We do have anaphora resolution code within opencog. It is supposed to do this kind of abstraction. It is located in the opencog/nlp/anaphora directory. @anitzkin when you say that the...

To be specific: when I last looked at the output of the anaphora resolution code, it was correctly handling both of the example sentences: `John steals the orange and eats...

The anaphora resolution code was built for the relex output because you guys were not ready with R2L at that time. Now that you are ready, it should be ported...

I mean, to be clear: I envisioned this block of code as the central place where we'd be carrying out not only anaphora resolution, but also the word-sense disambiguation, and...