[Biojava-l] The org.biojava.bio.program.* hierarchy

Thomas Down td2@sanger.ac.uk
Tue, 19 Jun 2001 13:28:00 +0100


I'm starting to get a little worried about the growth
of the bio.program hierarchy.  The original contract of
this package (and all it's sub-packages) was for wrappers
for interacting with external tools.  This seems to be
stretching thinner and thinner, though...

What definitely belongs there:

  - All the search-related stuff.  That's fine, and shouldn't
    move an inch.

Marginal cases:

  - GFF.  I think putting this package here was probably the
    root cause of the growth we're seeing now.  On the other hand,
    it's been there since BioJava 1.0, so it probably shouldn't
    move now.

  - DAS.  Interaction with external tools?  Possibly.  It seems a rather
    different case than running BLAST or FASTA, though.  Today, I'd
    probably just choose bio.das.

Things that definitely seem wrong to me, and might not be too
late to fix:

  - XFF.  This went into the `program' package by analogy with GFF,
    and also because it is used by experimental versions of DAS.  It's
    wrong, though, and I'm going to move it before 1.2.  What do
    people think of org.biojava.bio.seq.io.xff?

  - GAME.  I think this should move at the same time as XFF.
    bio.seq.io.game anyone?

  - Ragbag.  David's cool new sequence-assembly management stuff.
    I don't think this relates to external tools at all.  How about
    bio.seq.ragbag?

Sorry to make a fuss about this...  It's just that the BioJava
project now has rather a lot of packages.  We're already seeing
that new users often have some trouble `finding their bearings'.
So I'm keen to keep the package hierarchy as clean as possible...

   Thomas.