Latest modell runs (should have ingested some of the recon data!?): All dynamical modells (the GFDLÂ´s, HWRF, NOGAPS, except the GFS), show the nothward component to the westward motion to begin sooner, so that Ike stays more to the northern side of Cuba. The GFS sends Ike to the southern side. In the later forecast periond in the central gulf all modells (exept the NOGAPs, which send Ike straight NW to New Orleans) turn the storm from NW back to the WNW and send the storm to Texas. So Florida (except the Keys) is a bit out of the focus for the moment.
Is there a good source for the ECMWF-runs?
edit: Found a good source for the ECMWF with interesting comparison to GFS: ECMWF
So the ECMWF follows the general trend and sends the storm to texas.

Newest Center Fix: Nothing new. Data show a strong cat. 3 with 946 hPa central pressure. Movement straight west between the last two center fixes.