[Open-bio-l] LSIDs

[Brian King]

> the former or latter categories? I could resolve
> some
> LSID to the IUPAC (or whatever it is) table for DNA
> I
> guess. Different translation tables would also count
> as LSID-able but an entity representing the process
> of
> translation would not. Or am I missing the point
> again?

Matthew,

Whether something is identified by an LSID or not
depends on its persistence and its biological
significance, I think, so there are gray areas. 
Things identified by LSIDs have a format, so what is
the format of the IUPAC table?  Would identifying
formats by LSIDs lead to requiring formats of formats?
 Is a persistent, location-independent internet name
for the IUPAC table needed?

> And lastly, has anybody got their hands on a JNDI
> implementation that deals with URNs in a way that
> lets
> you plug in your own resolvers without writing a
> million JNDI support classes? Please?

I tested my LSID client using the JNDI implementation
that is built into Sun's JDK 1.3. It worked on Mac OS
X without writing any JNDI specific classes.  Are we
talking about the same thing?  Here is a code sample
form the API showing the use of JNDI.  This is an
abbreviated version of grabbing meta-data for a URN:

public Attributes getAttributes(String urn) throws
MalformedURNException, ResolverException
{
  // translate URN to LDAP Distinguished Name
  String dn = toDN(urn);

  // get directory object and return attribute
  DirContext lifnEntry = (DirContext) ctx.lookup(dn);
  return lifnEntry.getAttributes("");
}

Simple!

Regards,
Brian