For what it's worth, Mr. Sternbach's magazine article on the Constellation
class not only shows 'em having a 100-year regular service lifetime (heck, they put Stargazer
back into service after she was recovered in "The Battle") but suggests they will probably stay on ready reserve for a good while longer, needing only fuel and computer software and whatnot to be ready to go. Pretty impressive for a design Picard seemed to be suggesting was not particularly well engineered.
I never even thought the Constitution
class was necessarily out of service in the TNG era, although they were probably pushing it. One appears to have bought the farm at Wolf 359, judging from the debris; perhaps it was on training duty and relegated to intrasystem or local warp cruises or something.
Anyhoo, I don't like the idea that the Federation can build so many ships in a year that the Wolf 359 losses are rendered less significant than they came across on screen. The battle wasn't planned to be shown at the time, so the figure wasn't chosen because of budget constraints or anything, but is definitely an important clue to the overall size of the Starfleet. This is another thing that I plug in and keeps bringing me back to the 5000-8000 ships figure. Making of Star Trek suggests "hundreds" of ships in the TOS era Starfleet, if I remember correctly, so this is also useful to guide the thinking.
It isn't so much that this still isn't enough ships to allow for hundreds of Excelsiors
. Rather, the question is where we want them to be on the totem pole, so to speak. They still seem to be relatively large compared to a lot of the later designs, they appear to be to be multimission ships and used as such, and they are definitely still in frontline service, even for hazardous duty. I believe she was the fleet-leader of her day, and that building one easily takes the resources equivalent to four Miranda
-class ships and probably several years.
On the other hand, we must remember that the author's work supposes a notable modularity and solid warp dynamics for the basic design, which would encourage the building of more even a good while after the earlier runs, and the uprated version that appeared in Generations probably wouldn't have been worth doing for Starfleet if the class were small and intended to stay that way.
Thinking of the kind of resources that allowed a later Starfleet to plan for an initial run of six Galaxy
-class ships and six more in the freezer, and working back from there to what an earlier Starfleet would have tried for with its fleet-leader of the day and accordingly smaller resources, I feel like 50 or maybe 64 Excelsiors, something like that, shows a nice healthy run for a successful design--without seeming like they must have ditched everything else for a while, because as noted before, they seem to be in the business of trying what are presumably smaller runs of many different and specialized designs. I doubt that's something that only started in the TNG era.