Clark Lindsey has lots of links and thoughts about yesterday’s loss (not a permalink–just keep scrolling). I agree in particular with this:

As Elon Musk has said, expecting a rocket to work perfectly the first time is like expecting a huge, elaborate software program to run bug-free the first time it is run. (Fortunately for programmers, a computer usually doesn’t burst into flames if there is a fatal bug.)

This is why I and many other alt.spacers prefer space transports that can be tested incrementally rather than requiring “full up” performance the first time out. While the Falcon is partially reusable, it doesn’t offer the “fly a little, test a little” capability of vehicles with flyback stages. The SpaceShipOne test flights, for example, revealed several significant design problems but the increments in the system’s envelope expansion were always small enough that the flaws did not destroy the SS1. Instead the pilots were always able to return the craft safely and the particular problems found during a test were fixed before the next flight.

Yes. That’s the one thing that’s remained “old school” in SpaceX’s approach. You can test subsystems until the end of time, and still not know if the entire beast will work together perfectly, all singing, all dancing, the first time. NASA took this approach with the Shuttle, and got away with it, but they spent a hell of a lot of money on it. This is why it’s nice to have vehicle that’s not only recoverable, but one that lands the same way it takes off, so that it can be incrementally tested. Now that some other companies are starting to take that approach (both horizontal and vertical) it will be interesting to see how much more of an incident-free (or at least vehicle-loss-free) test program they have.

In the meantime, good luck to SpaceX. There’s no reason to think at this point that they can’t be as successful, ultimately, as their predecessors that cost much, much more to develop, but still had early failures.

[Update in the late afternoon]

Other people are having similar thoughts:

I guess that when you have an incredibly complicated system like Falcon or like other existing orbital vehicles, where everything has to work just right, there are almost no margins, and nothing can be flight tested beforehand, risks and sucky days like these are inevitable.

I’m glad that for our suborbital vehicles we will be able to do things like cutting our teeth on takeoffs and landings hanging under a tether. While we’ll still probably have our ulcer-inducing moments where we have to push the envelope into some new regime that we haven’t tried before, and where something could go wrong, those will be fewer and farther between. Trying to get every part of a rocket vehicle like that, with all the subsystems working perfectly from the start is a real challenge. SpaceX has a phenomenal crew, and I’m sure they’ll get this figured out, and probably make a whole bunch of money on this, but I’m glad that the approach they’re taking is not the only way to solve this problem.