NASA's new launcher strategy, which is slowly emerging on capitol hill, relies on a series of shuttle-derived rockets and an SRB-based "stick launcher" to send humans and cargo to the moon. It's likely that some of these shuttle-derived rockets will use a redesigned solid rocket booster, which is stretched to five propellant segments instead of four.
This booster was first tested in fall 2003. It adds more thrust and, more importantly, greater impulse than the traditional Shuttle SRB. However, the addition of a fifth segment makes the booster casing heavier, creating an unforseen problem.
When the shuttle SRB's splash down in the ocean, they impact nozzle-first. The boosters bob in place while divers attach a plug to the nozzle. The trapped water inside the booster is then purged until the booster floats horizontally on the surface.
With a five-segment SRB, the casing is heavy enough that the nozzle sits far below the surface of the water--so far, in fact, that the divers are prevented from attaching the plug according to OSHA regulations.
Maybe NASA can seek exemptions to the troublesome OSHA regs. Failing that, engineering solutions will be necessary. If NASA replaced some of the SRB's metal segments with filament-wound segments, that would reduce the weight. Filament-wound SRB's were developed during the 1980's, but were rejected at the time because the casings could be damaged upon impact with the water, and it was too hard to detect the damage to the structure.
A more involved approach would be the addition of a scissor-wing and control fins to the SRB's. The boosters would glide back to Kennedy Space Center instead of splashing down in the water. This would save time and money (no boats or divers to retrieve the boosters,) but would require a pretty penny to develop.
It would be a shame to let bureaucratic considerations sideline an important upgrade to the shuttle system. Then again, it could force NASA and its contractors to innovate.
EDIT (20 Oct 2005): It has come to my attention that a Wikipedia article is treating this post as an authoritative source regarding the five-segment SRB. I should warn the readership that my info comes from just one second-hand source. I can't vouch for the authenticity of this info because I haven't done the analysis, at least not yet.
In a larger sense, Wikipedia readers should be skeptical of everything they read on Wikipedia. Journalism is only as good as its source. Wikipedia's writers aren't always the most educated or unbiased historians out there, and even the good ones can fall prey to a bad source. Don't take my word for it--take Michael Isikoff's.