[Nitro] RDog ping

James Britt james.britt at gmail.com
Mon Jun 26 18:49:26 EDT 2006


Andrew Thompson wrote:

> Why use YAML at all? Why not database back this information? The
> original plan for rdog was to use Og to store all the information. The
> YAML stuff was only a transitional thing because we couldn't get Og to
> work. That has been resolved, but we never finished porting it to use
> Og.
>

In case my earlier posts were not clear, I see the YAML as an 
intermediate way to get the data into someplace for Og/Nitro to play with.

Yes, I imagine now that the hacked ruedoc code I'm using could just poke 
stuff into a database, by-passing the saving to disk/loading into 
Og-based object step.

> I have actually just checked out the repository and installed nitro. I'm
> gonna try and get rdoc-devel (my hacked rdoc) to work with ruby 1.8.4
> (there's probably rdoc changes again that are breaking things). I'm also
> gonna try and transition to a proper og generator and move away from the
> YAML one. I tried to kludge the Og one onto the existing YAML one which
> just added more issues than it solved.
> 
> I think YAML is not really suited to this anyway, it's harder to
> generate a list of all classes/methods, you often have to load more into
> RAM than you need and it's a pain to search. I think Og would do an all
> round better job.

Well, here's the thing: is it feasible to just keep the entire API in 
memory, using some data structure suited for searching and lookups?

E.g., using Madeleine.



-- 
James Britt

"The use of anthropomorphic terminology when dealing with
computing systems is a symptom of professional immaturity."
  - Edsger W. Dijkstra


More information about the Nitro-general mailing list