From what I've read SRS ballasts will not reliably ignite T8 lamps (if at all, most of the time only electrode glow is possible). Although there is probably little commercial interest in doing so, would there be a way of engineering an SRS ballast capable of igniting T8 lamps reliably and that could, in theory, be commercially feasible ? I am not talking about simply changing capacitor values in existing SRS ballasts, but starting from scratch where every variable can be tweaked (capacitor values as well as the number of turns on the choke, the diameter of said wires, core size, etc.). The distance from the grounded reflector to the lamp would also be adjusted if needed.

SRS from it's principle provide the input voltage for the lamp ignition, nothing more. It provide cathode heating controlled by the presence of the arc (no arc -> cathodes heated, arc present -> cathodes not heated), while it's design means no significant extra losses compare to the preheat series choke (the most efficient for it's size before electronic)
For T12 the 220..240V was way sufficient, so it was used for them.

For T8 the only way would be to operate it on higher input voltage, soou will need a step up transformer. But that would mean two consequences: At first it will be quite heavy (The SRS itself would have to be bigger, plus the size of the transformer). And as second, the SRS itself would have higher losses, again the transformer will add it's losses as well.

From that perspective, the RS concept used in the US will offer better efficiency.

But still, the T8 have another problem: Their hot cathode ignition voltage becomes quite close to the cold cathode one, so you will have it is not possible to choose the ballast OCV so, it will reliably start any lamp, while it still won't cold start them. F36T8 requires 800Vpk minimum required for reliable hot-start over the rated temperature and aging conditions, 600V maximum to avoid cold-starts), while with F40T12 it was ~280Vpk for a reliable ignition and at least 400Vpk to initiate cold starts, so a complete 200..280Vrms range mean it will be starting correctly.

Assuming, sol, that you are using the term SRS to refer to North American rapid start ballasts, as well, both electronic and magnetic RS ballasts already exist for F17T8 and F32T8 lamps.

@Medved : Thank you for the explanation. From what I gather, the T8 lamp's electrical parameters prevent such a ballast to be practical.

@TheMaritimeMan : I was refering to European SRS ballasts and 36 watt T8 lamps. You are right, magnetic rapid-start ballasts were made for F32T8 lamps here in North America.

For the European T8: Technically it would be possible to design a RS ballast for them, the eventual cold starting is not assumed as a problem anyway (most glowbottle starters preheat for so short time, they effectively do cold-start as well), but the RS would be way more heavy, complex and expensive than a series choke with a glowbottle for the preheat.

The US F32T8 RS do suffer from cold-starting, but the main aim there is just to start the lamp. The F32T8 on 120V would need step up transformer style ballast anyway (the functionality part making the ballast heavier and more expensive than series choke preheat), so the auxiliary windings to form an RS are then simpler and cheaper than the starter socket with the starter for a preheat.
And you may notice even in the US, than all lamps capable to operate with just a series impedance ballast (so sufficing with 120V OCV) are still mostly used in a preheat configuration, only the tubes, where the 120V OCV is not sufficient were mostly used with RS...