Arnon Dar (Technion) and Alvaro De Rujula (CERN) report:
We use the infamous Cannonball model (CB model) of GRBs
to predict the superluminal motion of the source of the radio
afterglow (AG) of GRB 030329 (GCN 1985, 1997), potentially observable
with the VLBA (see, e.g. GCN 2129).
The CB model has been succesful in describing GRB 030329 and
its AG, in predicting the supernova (SN) contribution to the AG,
and even the date at which the SN was to be convincingly seen .
In the CB model, the cannonballs responsible for the GRB and the
AG light move superluminally in the sky, as they travel away from
their projenitor SN. THIS MOTION MAY BE OBSERVABLE IN THE
RADIO AG, IF ITS LOCATION IS FOLLOWED AS A FUNCTION OF TIME .
The SN cannot be currently seen in the radio, so that it is the
motion of the radio-AG source (the CB) that one may try to observe.
In the case of GRB 030329 there were two CBs, one of which dominates
the AG after t ~ 1 day. Using the parameters of this CB determined
in  (initial Lorentz factor 1477, viewing angle 2 mrad, and
deceleration parameter 476 kpc) and Eqs. (30, 31, 41) of  we
obtain the angular distances "alpha" (in mrad) of the dominant CB to
the SN as a function of "t" (observer's time in days). For the current
cosmology (Omega_Lambda=0.7, Omega_Matter=0.3, H_0=75 km/s/Mpc; and for
z=0.1685, so that the angular distance is 0.64 Gpc) some representative
(t, alpha): (0,0); (1,0.22); (2.7,0.49); (5,0.74); (10,1.1); (30,1.9).
This means that from day 2.7 to day 10 the radio source may
have moved 0.6 mrad, and from day 10 to day 30 it may move
an extra 0.8 mrad. Such motion may be observable with the VLBA.
 Dado et al. astro-ph/0304106.
 Dar and De Rujula, astro-ph/0008474.
 Dado et al. 2003 A&A 401, 243.