But several persistent negatives of solids should be considered:
1) They are fabricated with their complete propellant load, and thereafter all handling, including any long distance transport, must haul the whole thing. Or as with Shuttle SRBs, chop it up into several lengths, each of which strains transport capabilities, then assemble it, which of course means vulnerable seams.
2) Once assembled, a solid rocket has a fixed, invariable thrust profile--ignite it, and in the best case scenario, it will burn in a known but unalterable fashion.
2)b)--Unfortunately it is not so easy to make that profile exactly known in advance; this was an issue with the Shuttle, since any deviation in profile between the two SRBs would throw unwanted yaw into the ascent profile, even assuming the average thrust would be close enough to plan. I suppose one could deal with small variations by steering with the gimbaled nozzle. But it meant that the SRBs had to be assembled from matched pairs of segments, which had been filled and cured in parallel so as to guarantee maximum identity of conditions.
3) Just as liquid fueled rockets are subject to "pogo," which is a manifestation of resonant coupling to acceleration causing the propellant supply to pulse, solids are subject to a comparable, analogous sort of resonant vibration--think of the huge SRBs on the Shuttle as a kind of giant whistle! One might say, oh, we know the length and diameter and fluid properties of the burning propellant, so just design for that--but of course the internal diameter and geometry is always changing due to the combustion of the material. Therefore the SRBs on the Shuttle were infamous for their powerful vibration giving the crew of the missions a rough ride--the early crews especially who remembered what Saturn V and other liquid rocket launches were like especially noted the unpleasant difference. Yet the Orbiter of course rode on its own liquid prop rocket thrust, somewhat insulated from the vibrations of the SRBs by two buffering interfaces, Orbiter to tank and tank to solids. Imagine how godawful the ride of the proposed Ares 1 of Constellation would have been, with the whole upper stack riding on top of one 5-segment SRB! I suppose pogo on the launches that suffered it (several early Saturn V launches were nearly destroyed by it, one was seriously deranged) was worse--but pogo was both intermittent and eventually manageable. The vibration of the solids is much harder to get rid of.
I am not such a big fan of solids, obviously! They are preferable in my view to hypergolic launchers. The actual economics remain murky to me as well--solids are often touted as cheaper. Well, it is certainly true that a given high thrust can be achieved with a solid engine that takes overall less development time and budget than making a liquid engine of the same capability. Shuttle SRBs put out considerably more thrust than Saturn program F-1 engines, but the latter was more difficult to develop.
If however we were to go over to trying to reuse the rockets, well, a solid rocket mainly is the propellant. It is not entirely the propellant; OTL it was supposed to be economical to recover the spent SRBs from a Shuttle launch, examine, refurbish, and refill the segments then reassemble boosters. Clearly the nose section, which caps the pressure load and holds the parachute, and most of all the nozzle section which not only has a carefully shaped nozzle but elaborate equipment to gimbal it, would be cheaper to reuse than discard, and I suppose that one should not underestimate the value and cost of the intermediate propellant segment exteriors--for these are pressure vessels, cylinders of considerable tensile strength that hold in the pressure of the burning propellant. Steel is not free! Well and good, but in fact many sources (anecdotal, I've never seen an actual cost breakdown) say that in reality, the additional cost of recovering the spent boosters, breaking them down for shipment, shipping them back to Utah to the Thiokol plant, refurbishing any damage or wear from the last launch and inspecting the units, all that versus the first cost of assembling new units from scratch, were so high that it would either have cost little more, or some say would actually have been significantly cheaper, to forego reuse, just let the SRBs splash and sink, and build new ones for every launch. Not 100 percent of that is inherent to the concept--some of it involved the refusal of Thiokol to set up operations in Florida which would have saved a lot of shipping costs. (But STS was meant to operate out of Vandenberg AFB in California as well; either Thiokol would have to have duplicate plant and workforces there too, or some means of shipping stuff from Florida to California and back again would have to be devised for those launches).
Had the decision ever been made to admit the idea of recycling SRBs was a dumb one, and the practice of retrieving the spent ones abandoned, not only would money be saved, but a simple modification of the SRBs removing the weight of the parachutes would presumably allow slightly greater Shuttle payloads too--the system took a hit when the slightly heavier more secure O-ring design was adopted after Challenger, and presumably would benefit from saving any weight on the boosters.
Clearly, although solids appear to be cheaper than a liquid design, if they really are cheaper is not by an order of magnitude. Whereas developing reuse even of just the first stage can drastically lower the overall costs of a launch, if the recovery and refurbishment process is not too expensive. These savings are not realistically realizable for solids.
Of course that does not matter at all if no one is attempting to recover anything and launches all remain expendable.
I have been writing as though the Shuttle SRBs are the only solids to ever be involved in launching which is absurd of course; the vast majority of solids lit for orbital launches have been much smaller units attached to smaller expendable launchers. It might seem that some problems especially highlighted by the huge Shuttle SRBs, challenged only by very large ICBM designs (or surpassed, in theory, by the hitherto yet to be proven 5-segment versions desired for Constellation and the SLS) might be mitigated by using lots and lots of small boosters clustered, but that raises additional problems too. Suppose that the two SRBs of the Shuttle had been replaced by 12 integral, unsegmented disposable solids adding up to the same reaction mass. (Why 12, rather than 8 for the 4 shippable segments? Because each SRB also included a nose and tail section, so it was 6 units shipped, not 4, for each--the segments were most massive to be sure, so perhaps we could get away with 10 separate solids instead of 12). Given the desire to put the Orbiter at the bottom of the stack and run its 3 SSMEs from the ground (necessary because of the difficulty of starting the sophisticated hydrogen high pressure engines, which would make air-lighting them a dubious risk) only a portion of the circumference of the propellant tank would be available, yet the nearest pair of solids to the Orbiter would be much nearer than on the OTL 2-solid design, exposing it to heat and closer vibration and sonic intensity. The less of the arc we use, the more asymmetrical the thrust, biased to the far side of the tank from the Orbiter and given the Orbiter's limited thrust especially at sea level, this would become impossible to balance pretty soon. One could go to an air-lit strategy and move the Orbiter up the stack, but it still has to be at the bottom of the tank or its exhaust is too great a hazard for the exposed part of the tank, and the vibrating and hard to control solids would become the only thrust source; also we'd want to cluster them tighter, which means the ones inside the cluster are surrounded by other solids heating and vibrating and sonically blasting them, which in turn affects reliability.
Most solids are not used for something huge like a Shuttle launch of course, which mitigates the problems! Generally speaking, the smaller the rocket you want, the cheaper and easier which is just common sense. An ATL where maximum single launch masses are kept modest could operate on, well, rockets we already had OTL (despite the hope the Shuttle would replace them). I dislike Titans (until they went back to kerosene-oxygen for the liquid core booster anyway) because I hate using hundreds of tons of hypergolic propellant even more than I dislike solids, and Titan used both, so I'd be rooting for Atlas and Delta upgrades, both of which have evolved toward what is routinely considered a "heavy" payload nowadays, around 20 tonnes to LEO. For these, modestly sized solids that can be built and shipped at reasonable sizes in moderate numbers are adequate, and give a standard booster core considerable range of upper stack mass hence payload. And if we liked we could evolve the core stages to become recoverable and reusable in various ways--simplest being to use a parachute to soften splash-downs and fish them out of the water with boats. (This works for Americans and the French launching from Kourou anyway--for Russians or people launching from the Australian or Algerian inland sites used earlier by Europeans, the spent stages come down over land, and to recover one I would seriously propose developing helium airships for the job; the biggest helicopters are rather small for the job!)
But it is certainly true that there are strong reasons why solid boosters have become very popular and ubiquitous in the launch business. As long as launch vehicles remain largely expendable, and modest in size, they are the smart way to go. You'll note I don't even object to their low ISP, as with boosters what matters more is thrust and density of propellant storage, and solids excel at those features, and along with them are unsurpassed for storability as well. Over decades, the grain of solids will deteriorate, which is a big issue for say the Minuteman missiles, but presumably no national or commercial launch program will allow useful boosters to sit too long before being used.
