parting thoughts

[Jason Bell <bellj@graphics.lcs.mit.edu>]

BMG group:

here's a list of TODO's, etc. that I thought of at sort of the "last
minute."  I'll be around (via email) to help with the project for the
next few months, so feel free to email my graphics account.  If for some
reason I'm not answering, send me a line at jbell9999@yahoo.com (but I
like to keep the graphics stuff separate).

BMG stuff to be done:
Ground cover information - this should be integrated into the pipeline.
This should also be "spit out" in LocationServer compatible data files
(Patrick knows the format).

Building transformations - the building transformations should alse be
integrated into the pipeline.  it would be good if there were transforms
embedded into each floor data file, but a single master file is okay
too.  The one that exists on /scratch/fake_data was hand-culled from a
file I found in Sean's old directory (hence the name fake_data).

Errors in Portals - some Portals are still not being converted
correctly.  Use the RouteFinder application to see where the errors
are:  it's good for regression testing

LocationServer stuff to be done:

XML - the network protocol should be changed from my hacked-together
definition into some sort of well-defined XML or something.  Java RMI
_could_ work if there's an easy, cross-platform way to serialize and
unserialize their datatypes.  But right now, were stuck with using Java
clients.

3D polygon information - we still can't serve it.  Patrick - you should
add some methods and stuff to LocationServer to do this

Elimination of virtual tags - Lee's project makes this feasible.  As
long as all the servers and clients are "speaking" the same Space name
information, we can allow Lee's virtual tag database to be separate.  Of
course, LocationServer only serves geometry information in this case.

okay.  any questions, just drop me a line.

-Jason