[Biopython-dev] record iteration

[Andrew Dalke]

Me, on standardized Martel tag names.

Jeff:
>This all looks good.  Do you have a sense on whether there should be a
>unique prefix or suffix to indicate that a standardized name is being
>used?

Well, it's XML so the best solution is probably XML namespaces.
They would look like this:
  biopython:sequence

but I don't know enough about them.  I've only been doing
non-namespace work.  To make things more fun, the SAX 2.0 API
is slightly different for namespace tags as compared to
non-namespace tags, and I don't know how they are supposed
to work.  The documentation I have is pre-2.0, and my
"Python and XML" book hasn't arrived yet.

> e.g.  _m_dataset or something like that.  Since these names are
>pretty common, especially in this domain, it might be easy to use a
>standard tag name when none was intended...

My plan is several-fold:
 - make up my own tag names, since it's just us for now
 - document when/how they are used
 - look for existing tag names (BTW, I don't understand GAME
     all that well)
 - convince others to use Martel, or at least the ideas behind
    Martel (if no one else uses it, there won't be a namespace
    problem)
 - keep going blithely until there is a problem and/or until
    someone tells me how to use SAX 2.0-style namespaces
 - present the general Martel/parsing idea in my Biopython talk
     at the O'Reilly conference
 - bring up the specific problem in a lightning talk
 - convince others at the hackathon to help me, as I don't have
     enough breadth to get things right.

Going to be a busy January. :)

                    Andrew