Off Armageddon Reef opens with humanity losing, and I loved it for that. Not losing a battle. Losing the war, completely and permanently, over decades, to an alien species that doesn’t negotiate and doesn’t innovate and simply grinds the Terran Federation down until there’s nothing left. It’s a genuinely bleak hundred pages and it’s the best thing David Weber has written outside the Honorverse. It’s superb.
I’ve reread this one several times now, which is not something I say about many doorstops. The concept is what keeps pulling me back. One colony fleet escapes by faking its own destruction, reaches a world ten thousand light years out, and then the administrators in charge make a decision I’ve never fully made peace with: humanity must never industrialise again, because industry has an emissions signature and the Gbaba follow signatures. So they wipe the colonists’ memories, tell them God made them a few days ago, write themselves into scripture as archangels, and build a church whose entire purpose is to make sure nobody ever invents anything.
The setup is the standout, and it isn’t close
The Terran Federation and the Gbaba are what I’ll remember from this book in ten years. What gets me is the sadness of it. Humanity takes the fight to Gbaba worlds, wins engagements, does everything right, and it doesn’t matter, because the enemy replaces losses and humanity can’t. Weber spends real pages watching the species lose in increments, and I felt every one of them. It’s genuinely grim.
I’ve seen the starting-over-on-another-planet idea done several times and done well. Doug Dandridge’s Exodus is a good example of the same instinct, a fleet running from something unbeatable and trying to build a future somewhere else. What Weber adds is the twist of deliberate stagnation, and that’s where the book stops being comfortable.
I found it genuinely hard to swallow. Millions of people held in the Dark Ages, forever, for their own protection, with a religion built specifically to punish curiosity. Pei Shan-wei’s faction wanted the knowledge preserved and released when it was safe, and Langhorne killed them all for it by sinking an entire landmass from orbit. He’s the villain of that exchange and he might also have been right, which is the uncomfortable bit. My own position lands in the middle. Hold them down five hundred years, let the Gbaba lose the scent, then let humanity start learning again. That seems obvious to me and it’s what neither faction in the book was willing to do.
There’s a hole in the premise that I want to flag, because it’s the sharpest criticism I’ve seen levelled at this book and I think it’s fair. The Federation could copy a human mind into a PICA. That technology existed. It was apparently used for recreation, so people could go rock climbing in a body that wouldn’t die, while humanity was simultaneously losing a war of extermination for want of good people. Nobody thought to make twelve copies of their best physicist. Nobody made twenty of their best admiral. The book’s answer, as far as it has one, is that it just didn’t occur to anyone, and that’s thin.
Does it sink the book? Not for me. It’s a real hole and I can live with it, partly because the story Weber wants to tell needs Nimue to be singular and unrepeatable, and partly because “nobody thought of it” is depressingly plausible for a bureaucracy in a losing war. It’s the kind of flaw that bothers me more when I stop and think about it than it ever did while I was reading. But it’s there, and if it’s the sort of thing that ruins a premise for you, you should know going in.
Characters I actually got attached to
Cayleb and King Haarahld are the two who got me. Haarahld especially. He’s patient and sharp and he’s the first person to work out roughly what Merlin is without being told, and Weber kills him at the end of the book, which I think is the strongest decision in the whole novel. Cayleb goes from a likeable prince to a king in the space of one battle, and the film of that transition is convincing.
Merlin himself is interesting for reasons the plot doesn’t spend much time on. Nimue Alban died covering the evacuation eight hundred years ago, and what wakes up in the cave is an android carrying her personality and memories. Whether that thing is still Nimue or a very good recording of her is the only genuine interiority in the book, and it’s mostly left sitting there. Her decision to present as male, because a woman won’t get a hearing in a medieval patriarchy, is a practical choice the book treats practically. I’d have taken more of that and less of the cannon.
The odd one out for me is Erayk Dynnys. He’s the absentee Archbishop of Charis, corrupt and weak and not remotely admirable, and I found myself drawn to him anyway. He’s the first real sign that Weber isn’t going to write the Church as a monolith of villains, and there’s something in a man who’s been coasting his whole life suddenly having to look at what he’s part of.
The scattered use of lost technology through the book is a real plus too. It’s never a magic wand. It’s a cave, a library, an armoury, and a lot of restrictions, and the fun is watching Merlin work out what he can push through the gaps in a scripture that banned specific things and never thought to ban the rest.
The naval writing is exceptional, and there’s too much of it
The descriptive writing on the space battles and then the naval battles is genuinely superb. Darcos Sound is the payoff, where a small force of galleons annihilates a much larger galley fleet because galleys simply cannot answer a broadside, and the whole sequence is clear and tense and horrible. I found myself wincing for crews who never even got names. Reactor’s write-up called it sea battles and grand emotion, which is about right for the last third of the book.
The rigging of the ships is described exceptionally well. I did not need to know how ships were set up on every other page. That’s the trade with Weber and it always has been. He’ll spend twenty pages inside a character’s head on the development of a cannon, and about six of those pages are interesting. Six. The middle of this book sags badly under it, and the climax earns it back, and I understand why he does it without wanting quite so much of it.
The other thing that keeps throwing me out is the names. Cayleb, Haarahld, Zhaspahr Clyntahn, Erayk Dynnys. The in-world justification is that spelling drifts over eight centuries, which would be fine except it only ever applies to people’s names. Everything else on Safehold is spelled normally. I got used to it eventually. The first third was a fight.
Four out of five, and one warning
Good book, engaging story, and I’ve gone back to it more than once. That’s the real test. Four out of five.
There’s a lot of self-reflection in here that I could have done without. Characters thinking about what they’re doing, at length, repeatedly, and it’s a real waste that extends these books as a whole. If you like Weber you already know this and you’ve decided it’s worth it. If you don’t, this is the thing that will stop you.
The warning is about the series rather than the book. Don’t drag it out. If it takes a dozen novels to get Safehold back to twenty-first century technology, I’m not going to be along for the ride, and I say that as someone who’s just told you he’s reread the first one several times. There’s a version of this story that runs four books and lands hard. I’d like that to be the version I get.
If you’re coming to Weber for the first time, I’d still point you at Honor Harrington and On Basilisk Station before this, mostly because it’s a leaner introduction to how he writes. But if the premise here appeals at all, start here instead. It’s a better idea than anything in the Honorverse. And if you want the other famous novel about a small group engineering a civilisation’s recovery on a schedule, the Foundation books are the obvious companion piece, with the crucial difference that Seldon’s plan is designed to shorten the dark age and Langhorne’s is designed to make it permanent.