[Objcryst-foxx] Fox v1.5, new (last) beta, winXP status

[Favre-Nicolin V. <vi...@us...>]

	Hi all,

   I've had a few more reports about winXP crashes with the 3D display in F=
ox.=20
It turns out that all those who reported problem (so far) use an "Intel=20
Extreme Graphics" video card. I've played a day with it on a borrowed lapto=
p,=20
and it turns out many other programs (including very simple ones) using=20
OpenGL crash with this card when 3D acceleration is turned on.
   The bottom line is that I now label this as a bug in the 3D video driver=
,=20
unless someone has the same problem with a different video card. The=20
workaround for this is to diminish hardware video acceleration, until=20
Direct3D acceleration is disabled. This can be done in the display properti=
es=20
(right-click on the desktop to access display properties and navigate to th=
e=20
advanced settings till you can chose the level of acceleration).

   Otherwise, I've uploaded a new beta version of Fox, which should be the=
=20
final 1.5 release. This version changes:
=2D avoid crash under windows when there are some points with null intensity
=2D automatic resizing of panels, for people working with really large=20
structure(s)
=2D the cimetidine example with extracted refection intensities has been up=
dated=20
with constant weights, so as to run in un-weighted mode (since the extracte=
d=20
intenisties in that file have uncorrect sigmes)

   The new beta can only be found from the Universit=E9 de Gen=E8ve:=20
ftp://ftp.unige.ch/pub/soft/crystal/fox

   There are a few things I'd like to add but it will have to wait for the=
=20
next (minor) release:
=2D Fox is very slow when updating the display of large structures (e.g. >1=
00=20
atoms in a Molecule), or converting them from Molecule to ZScatterer
=2D allowing the Molecule to have several levels of flexibility : (1)=20
rigid-body, (2) rigid-body with user-chosen free torsion, and (3) automatic=
=20
assignment of free torsion. Currently (3) is implemented, and I believe wor=
ks=20
relatively well, but it may not be best for large structures with few degre=
es=20
of freedom (you have to add a lot of restraints manually).

   Please tell me if you have any problem with these packages. I'll make th=
e=20
release in a few days.

	Vincent
=2D-=20
Vincent Favre-Nicolin
Universit=E9 Joseph Fourier
http://v.favrenicolin.free.fr
ObjCryst & Fox : http://objcryst.sourceforge.net