[Proteopedia] NMR format, bot tasks

[Ralf Stephan]

Let's tackle that first idea first.

Eran wrote 
> That's a really great idea if it can be done well. How would you suggest 
> doing this (of course automatically)?  Let's say you're on the page for 
> 1ea5, do you want to list, on the page for 1ea5, all other pages that link 
> to 1ea5?  Or all other pages that link to 1ea5 but are not PDB-entry pages? 

I want all those pages where that structure is displayed, the
algorithm would be like this:

- define a container holding a data structure with 1. protein
  structure names 2. list of strings (later holding the topic
    page names). The container key is 1.
- for all topic page names
-- load the page
--- for any scene or STRUCTURE statement, append the page name
    to the appropriate data structure in the container
- for all data structures in the container having names appended
-- load the structure page
--- complete the entries there (should be agreed upom the format
            first!)

>  I'm just worried that such an automatically generated list might, in some 
> instances, be long and irrelevant.  Of course as you've said, the person 
> who created the Topic page should link from the topic page to the PDB entry 
> page and vice versa, but they will sometimes neglect to do this.

I just see that PDBsum links to Proteopedia structures but those who
follow that link will not benefit from the work done on the topic pages.

> Good to know! Let's brainstorm this idea a bit further before talking about 
> implementation. As Jaime said, there are currently no user-owned bots on 
> Proteopedia, so, at least for now, we would run your bot for you.

Jaime and me are already in discussion about details. I would have to
think about how I would test it here, but see no big problems.


ralf