Novice Timing Belt question

I have pretty much given up on trying to economically source a ball
screw for the long axis and am looking at timing belts.

From the CNC machines on the web it appears that the most popular
configuration to use is the fixed belt method ie "....the driving
axis, idle rollers, and platform proceed as a unit along a length of
belt that's held fast at both ends. Belt and drive sprocket impel the
load while linear bearings support it"

My question is why isnt the method where the belt is stretched around
a driving and idler sprocket used more often? ie "....The driving
pulley pulley and idler pulley of equal diameter are offset. The
load is linked to the belt and moved parallel to the span between the
pulleys. This system contains a platform or slider that moves on
linear bearings"

For diagrams of both systems see:
http://apps.mectrol.com/applications/doc/linear_day2.pdf

The reason I ask this question is that with the fixed belt method the
weight associated with the motor, idler rollers etc must all make the
trip around the table meaning bigger motors may be required, more
inertia etc and also the fixed belt method appears to require more
parts. I'm assuming that there are advantages of the fixed belt
system over the closed belt which more than offset these
disadvantages.

Thanks for any responses, in advance.

Robert