(I'd like to thank the Space Frontier Foundation and NASA for giving me this opportunity to present Space Access Society's position. The opinions that follow are not necessarily those of the Foundation or NASA – although they should be).

Henry Vanderbilt of the Space Access Society.

X-vehicles are hot in the space business these days – fashionable and fundable. It should come as no surprise then that all sorts of people are tagging their pet projects "x" in hopes of jumping onto the funding bandwagon. But while federal funding for space X-vehicles is available, it's far from unlimited. We have a strong interest in making clear what is and isn't actually "X", what we will and will not support the government doing with the available funding.

One pseudo-X example is the X-38 ACRV, a "Y" vehicle in X clothing, a routine operational mission-flying vehicle project disguised as an advanced experiment. A variation on this theme is "x" projects where operational mission requirements are mixed willy-nilly with experimental goals, as with the original X-34. Such confusion between experimental and operational goals fed to the original X-34 project's demise, and causes ongoing problems for X-33.

Y-vehicles are prototypes of ships intended, with minor production refinements, to carry operational payloads and perform operational missions. (Repeat after us: "Prototypes" are NOT X-vehicles.)

OK, you say, so how would we define a genuine X-vehicle? We're glad you asked... New aerospace vehicle technologies can be taken only so far in computer simulations and wind tunnels and test stands. There comes a time when the only way to pin down the remaining uncertainties is flight test – the sims and ground tests are good and getting better, but there are always conditions the sim only approximated, interactions the ground-testers didn't anticipate.

The wrong way to flight-test new technologies is to bring together a whole bunch of them directly into a project to build, say, a prototype airbreathing-to-orbit spaceplane (NASP) or (hypothetical example of course) an SSTO replacement for NASA's Space Shuttle. In theory this approach saves time and money – skip all the intermediate flight-test data-gathering and debugging, and go straight to a prototype as close as possible to the final operational vehicle.

There's a problem with this approach: the relatively large remaining uncertainties in the partially tested new technologies force the designers to use large safety margins, because the resulting vehicle MUST work, reliably over many flights – it's costing billions, it has a high political profile, and it has missions it MUST fly. High risks for high payoffs are not allowed.

The large margins translate to heavier subsystems. The heavier subsystems multiply more than add: heavier tanks require heavier support structures require more powerful engines require larger tanks require... More expensive materials and more exotic manufacturing techniques are dragged in to try to contain the weight increases. The vehicle size and cost balloon, the project bloats and stretches out, the final result is at best a marginally operable kluge.

As we said, a hypothetical example only. The right way to flight test new technologies is, well, to flight test 'em. An X-vehicle is an ad hoc flight demonstrator, designed to find out as quickly and cheaply as possible what happens when one or more new technologies are pushed to their limits.

X-vehicles can range in scope from a new rapidly solidified – unobtainium TPS sample bolted onto a sounding rocket for a few millions, to a package of mostly existing plus a few new technologies bundled into an integrated flight test vehicle for a few hundreds of millions. Either way, X-vehicles have no missions but building experience and returning data, and no payloads but instruments and in some cases pilots.

X-vehicles are essentially disposable – you don't waste resources on production – engineering, you don't include much systems redundancy, you build several copies and count on breaking one or two before the test program's over. After you've built and flown an X-vehicle, THEN you have the data and experience to design an operational prototype.

Paradoxically, doing two design-build-fly cycles, X-vehicle then prototype, historically ends up quicker, cheaper and more effective than trying to compress the process into one giant leap from the ground test labs to an operational vehicle.