Besides the seeing aspects, the concept of a co-rotating building utilized for the MMT was very practical for telescopes with a altitude-azimuth mount. Thus it was also considered in the early 80s by ESO for the New Technology Telescope (NTT) 3.5-m project. The NTT project team had some critical reflections on the MMT building and looked for further improvements: it was argued that if the MMT could improve seeing with a large frontal opening, adding another one at the back to flush the telescope with the wind flow might improve it further. At the same time it was still considered that smoothing a few edges would improve the quality of the airflow on the telescope.
Figure: The first version of the NTT building.
Figure: The final building of the NTT
Therefore the initial concept for the NTT building
was a mixture of old and new (fig. 
):
the building was basically a cylinder topped by a hemisphere,
chopped in order to have plane walls at the front and back and a
flush-through slit.
A model of this building was tested in the wind tunnel and the measurements showed that strong, high frequency turbulence was created inside the building by the wind going through the slit, with a peak frequency near to the first eigenfrequency of the telescope. This would have degraded the guiding performance by more than any seeing improvement. Still, a configuration with both ends open remained attractive because it allowed the exchange of the whole volume of air in the dome within a few seconds.
The designers added a semi-permeable windscreen at the front (see
fig. at page )
and closed the back wall with adjustable louvers.
Then, realizing that a
wind shielding system would be necessary independently of
the external shape, they abandoned the pretense of an
aerodynamic shape and redesigned the building geometry solely
on the base of its functional requirements - fig. .
A new, more extensive series of wind tunnel tests were conducted
with the new configuration (the main results of which are included in
chapter ), which confirmed the need for a frontal
windscreen with a permeability of about 20%. Experience on the
telescope, in service at La Silla since 1989, has showed that the
windscreen needs generally to be raised when the wind exceed 6-7 m/s.
In conditions of light breezes (2-4 m/s) without windscreen and the
louvers opened, the NTT appears to have virtually no local seeing.
However when the wind is stronger and ventilation must be reduced,
the seeing effect reappears.