Re: knngist plans - Mailing list pgsql-hackers

Oleg as I said I do believe some things could get better in docs.
Here are 4 main drawbacks in my opinion:
1. The docs lack example(s). There are some references to gistproc.c etc but this is not part of the docs and if I am someone that has just started this is hard and confusing (I shouldn't have to read the full technical solution to another problem in order to solve mine).
2. The 7 methods are poorly explained, at least theoretically. For example when I started I couldn't understand where each function was used. I had to read papers to understand the logic.
3. Describe situations that gist can be used. I am aware of spatial indexing because I worked on it, but I have the impression that there are more (e.g. computational biology).
4. Why not add some images/figures? (1 image = 1000 words)

GiST is a very important and useful part of postgresql but it's hard to get started with it. Moreover, I am sure it has capabilities, other than the profound, that are not known to many people (including me).
Plus knngist will give gist a boost, and I suggest that it shouldn't be delayed any more.

We have some documentation already, Teodor will post soon. 

I would really like to see what Teodor has done.

I could also prepare a document with the additions I would like to see in docs. Shall I do it? Of course it might take a while to finish it.