Following on from yesterdays landing failure, a few thoughts on how the data gathered from these flights could be put to use and what features the next version should have.
First of all, it should be designed for increased ruggedness and for reusability from the start (natch). This should mean an aim for aircraft like performance and an eventual aim of 'gas and go' operations. These vehicles should undergo an incremental test programme to determine reliability before putting on second stages and payloads. The staging and mass ratio between first and second should be optimised for reusability, if they find that the GTO missions put too much stress on the vehicles or don't have the margins for landing then keep flying them as expendables and only reuse for low orbit missions.
Second, I think should be a wider, shorter vehicle. Maybe around 6 metres wide first stage (compared to the current ones 3.8m) This is to increase stability on landing and to allow a larger fairing to be used. I would also attach the fairing directly to the first stage and put the second stage and payload (for unmanned launches) inside, a la Antares and Delta heavy. This way the fairing can be recovered with the first stage (I'm imaging it opening to deploy at separation a bit like the rocket in Thunderball). Another advantage is that larger, Raptor powered upper stages could be used on the heavy version. These could then lead to a reusable upper stage (one idea I have is a conical stage that incorporates a Dragon capsule for manned missions)
Finally, the 6m tanking can the used for the next generation raptor powered first stages, which could be built in various lengths and with different numbers of engines to replace F9 and heavy prior to MCT launchers coming online.
Of course, it is most likley the announcement in September will render these points all moot. Or I could have been prophetic....