Ian,

Looks like we both recall correctly but incompletely. The AAIB report of
the 1996 accident, at
http://www.aaib.dft.gov.uk/cms_resou...pdf_502118.pdf states:

"Several witnesses formed the opinion that this launch was faster and
less steep than normal, the launches already being considered, by
several witnesses, as being fast that day. One witness reported the
presence of a strong wind gradient, the surface wind being 5/10 kt. At a
height variously estimated by witnesses of between 600 and 700 feet the
glider was seen to 'speed- up' and its climb angle reduce. At about this
time, and whilst still in a nose high attitude and connected to the
cable, several relatively rapid oscillations in pitch occurred. One
witness estimated these to be as much as 15°, with the wings being seen
to 'flex' correspondingly an abnormal amount. At about this point,
the airbrakes were seen to briefly deploy. Almost immediately, they
deployed again, this time coincident with the right wing failing in an
upwards and rearwards direction, pivoting about its root end and
releasing a cloud of debris."

A doubled weak link was used, which could have allowed excessive loads
on the airframe in previous launches and in this one. Part of the
conclusion was:

"In view of the fact that the corrosion had developed over a long period
of time, during which the glider had been launched and flown without
incident, and that the aircraft had been launched by winch earlier on
the same day in similar conditions by the same pilot, it is considered
probable that excessive loads were induced on this occasion. This, in
turn, exploited the degraded strength of the wing spar, resulting in the
upper spar boom collapse in the right wing."

I read this as saying that continued overstressing on a weakened
structure led to the failure, with that days's overstress being the
final straw. The report says that other gliders flying from the same
airfield had probably been overloaded. Only this one failed
structurally, because of the corrosion.

As I wrote before, flying outside the envelope takes one into uncharted
territory. However, overspeeding on its own does not necessarily do so -
it depends on how the pilot flies the launch and when the launch is
abandoned.

To summarise how I understand the UK training:

1. If overspeed is clearly excessive from the outset, release and land
ahead if possible. If not possible to land ahead, see (3).

2. If moderately excessive in the lower 1/3 of the launch, give a chance
for it to settle down and if not, signal too fast.

3. If it doesn't settle down, reduce the load on the airframe through a
shallow rate of climb and release once a safe height is achieved.

4. Overspeed in the top 1/3 of the launch is potentially dangerous as
the glider is under its highest load at that point, and cumulative
overstress can lead to failure particularly if the structure is already
weakened.

5. Too slow is more dangerous than too fast, provided you fly the fast
launch so as not to overstress the airframe.

I'd add that all suspicions of corrosion, overstress through excessive
g, etc. need to be checked out by a qualified person.