[redland] Redland News 2000-03-27

[Dave Beckett]

I've been working on fixing Rapier's bugs after I rolled out  the
parser test suite:
  http://ilrt.org/discovery/swsw/pt
shows how 'alpha' it is

I've also been updating Redland and adding to the API such that
moving around the RDF graph is more natural, you can now use
arcs_in/arcs_out (and tests) on the model.  If you look at the
Mozilla interface IDLs, you'll see what I'm aiming at

Since working on several large RDF graphs indexed with Redland via
BerkeleyDB (henceforth BDB) I've decided to change the indexing
method to match what is actually used, as well as allow more
flexibility.  Where large is around 1.5M statements in 3 BDB hashes,
on disk size 2Gbytes.

It has also got to that stage where I recognise the need for some
refactoring of bits, presently I'm considering making the iterator /
stream classes be the same, and a more powerful data flow and
synchronisation channel.

But the goal for doing this work is to get some sort of query
implementation going, at least in a simple form (conjuctive joins,
something like Squish).

And I *might* have got somebody to do a MySQL storage backend, but
we'll see.

Dave