Your question can be broken down into three parts and you don't seem to consider the last issue:

• Is the pre-flash really necessary?
• Are old flashes capable of doing a pre-flash?
• Can the camera order old flashes to trigger with a specific power?

So let us go through them one by one.

Is the pre-flash really necessary? I have never actually measured, but the explanation sounds fair. With old TTL-capable flashes, the camera measured the light reflected from the film surface during the exposure and quenched the light strobe when enough light had been measured for a correct exposure. Allegedly, the sensor surface does not reflect light in such a manner, that the reflected light can be measured and used for flash metering. The problem is not really the sensor surface, but the antialiasing filter placed in front of the actual sensor. For obvious reasons, the antialiasing filter is built to let most light pass through and not reflect much light at all. The filter also does not behave the same in both directions, so if light passes through the filter and is reflected by the actual sensor surface, the light will not pass back through the filter the same way as it entered. It is therefore at least reasonable to believe this explanation.

Are old flashes capable of doing a pre-flash by having the camera using the available control interface to instruct the flash to fire twice shortly after eachother? Perhaps, not but likely with all flashes. I have used older flashes, which even when triggered at the lowest possible power setting, need a noticeable amount of time before being ready for the next flash. We are not talking about a very long period here, it is likely just a fraction of a second, but still too long for an acceptable shutter release delay when you actually want to take pictures.

The most important problem however, is that even if the camera could lure a pre-flash out of an old film-TTL-flash, the old TTL-systems would not allow the camera in advance to instruct the flash to use a specific power setting. There are likely differences between the manufacturers in the details of the interface, but old TTL-flashes were only controlled with 'on' and 'off' triggers and the correct exposure was achieved by controlling the duration of the strobe. After opening the shutter, the camera would signal the flash to turn on, start measuring the exposure and as soon as enough light had been detected, the camera would trigger the 'quench' pin of the TTL interface, causing the flash to extinguish. Also the newer 'digital' TTL systems differ between the manufacturers, but the principles are similar. The Nikon iTTL protocol operates e.g. with 127 power levels, so instead of just grounding the X pin, signalling the flash to trigger, the camera uses a serial interface to tell the flash e.g. 'give me a flash with 43/127 of maximum power'.

One might believe that it might be possible to control the power of the main flash by calculating the proper timing. E.g. if the pre-flash is 1ms and the camera calculates that 6 times the amount of light is required for the main flash, why not simply use the old TTL-interface to trigger a 6ms long strobe? That won't work, since the intensity of the light varies by time. The bulb needs a small amount of time to reach maximum intensity, the intensity will decay over time as the capacitor is discharged, and even after quenching and interruption of the current flow, the bulb will still glow for a short period. There is hence no linear relationship between flash power and flash duration.