BOMBERS:
Currently, I can think of no way a scramjet can ever replace the B-52 bomber (or any other offensive military aircraft, for that matter). Why? There have been documented cases of high-speed bomb drops where the bombs, instead of dropping out of the bomb bays like they were supposed to, instead hit the high-speed air layer flowing along the skin of the aircraft and bounced back into the bomb bay (and this was on EXISTING bombers). Just think of what would happen with a bomber that cannot travel slower than Mach 6 when it tries to launch a bomb. Just the turbulence created by breaking the smoothness of the aircraft's skin might be enough to wreak havoc on its payload.
External hardpoints are quite possibly also out of the question even if they could withstand Mach 7 air resistance, since in the moments just after release, the air layer running over the hardpoint would break, creating turbulence, and make the payload very difficult to control.
PASSENGER FLIGHTS:
Also, I don't think passenger flights would be at all economical with scramjet technology. Look at the size of the Pegasus rocket in comparison to the scramjet.
Let's make some rough ballpark calculations:
* The X-43A is a couple of feet shorter than a Dodge Viper (check it if you don't believe me)
* Now, scale it up to twice the size of a Learjet, so maybe more than a dozen people can fit in the passenger area.
* Similarly, scale up the booster (currently 17 m long). Assuming we're going to have great leaps in rocket efficiency, cut the proportional size in half.
* A Learjet 35A is about 48 feet long, which translates to 14.6 meters. So, we have a 30-meter-long scramjet, strapped to a booster that's now about 73 meters long, or three-quarters the length of a football field.
* This booster would also need to burn longer, since less acceleration = smaller g-forces on passengers, AND it would need to lift off from the ground. So, you'd actually need a BIGGER booster for passenger flight, even if you reduced the size a tad by making it air-breathing.
* Conclusion: To ferry the same number of passengers as a Learjet, you'd need a two-stage aircraft the length of a football field.
Is that more economical?
Addendum: Also consider that the current setup allows hypersonic flight for a few seconds. Any flight that'll draw customers towards hypersonic flight and away from conventional flight will need to be at least a few hours long. Tack on some more volume and weight for all that extra fuel.
Thus, I think that scramjets have the most promise as globetrotting cruise missiles or reusable spacecraft launch platforms. (Actually, using them cruise missiles probably wouldn't too practical, either, but I won't go into that unless by request)
Imagine a small-payload two-stage space shuttle replacement. Let's say the first stage (the booster) is also reusable. Conceivably, a size range similar to that described above for passenger flight could replace the current space shuttle and smaller rockets, lobbing small-to-mid-sized payloads into orbit. This is actually the only practical use I can think of for rocket-boosted scramjet craft...
...though it'd be pretty damn cool if they actually built a scramjet passenger craft.
GIANT MAGLEV LAUNCHER:
1. Maglev track (a la Shanghai Maglev train)
This is all well and good, but do you remember the massive hurdle the British supersonic car had to overcome? The sound barrier. Breaking the sound barrier with a tiny air layer beneath the car caused an extremely high-pressure wave to travel below the car, threatening to force the car airborne. I actually still don't know how they overcame this, but imagine the same thing but without such a high degree of aerodynamical optimization. (because it's not physically possible, due to the mechanism needed to constrain the projectile to the track) To put it mildly, the engineering involved would be nightmarish.
2. "Giant railgun" (a la BFG)
If a scramjet craft were magnetically launched in a fully enclosed barrel, the track-attachment problem would be solved, since the projectile would just float in the center of the barrel. The actual construction of a 10-mile-long magnet-laden tube (let's assume that this is sufficient) isn't that farfetched, since CERN's Large Hadron Collider is 26.7 km in circumference.
But, there is still the problem of boring a 10-mile-long giant tube AT AN ANGLE into the ground. The only equatorial place I could think of where the logistics wouldn't be staggeringly mind-boggling would be in the Andes. Boring the tunnels into a mountain has a few benefits:
a. Maintenance and rubble-removal tunnels could be bored in the side of the mountain, shortening them.
b. The mouth of the tunnel would be at a higher altitude, meaning thinner air, meaning less air resistance for the projectile to overcome during flight, meaning less energy needed to fire the launcher, meaning cheaper launches.
I think this is a neat idea for scramjet launch assists, acting like a hypersonic catapult.
In summary, I think that scramjet technology would not be economical or practical for currently envisioned application outside of spacecraft launching, and I think that both chemical boosting and maglev catapulting are relatively viable methods.
P.S. - Then again, the scramjet's minimum operational speed would increase on its trip to space, as the air would become thinner, requiring higher compression ratios for scramjet ignition.
Just my two cents.
