Technology

Catastrophic Failure in Fold Transit

fold-drive
fold-transit
safety-corps
certification
standards
the-institute
wrath-of-the-dragons
boundary
loss

Nothing has ever been lost inside a fold. Things are lost at the two ends of one, which is a different problem, a duller one, and — the Safety Corps would insist on this — a solvable one.

Higher-dimensional folding is geometrically lawful. The channel draws distant regions into adjacency, passes matter through stable aperture conditions, and releases. Nothing in that sequence is hazardous in itself, and the loss record contains no case of a vessel destroyed by the fold as such. The failure surface is the aperture — its formation, its stability while a hull is inside it, and its collapse — and every catastrophic loss on record is a failure to hold one of those three within tolerance.

This matters because an aperture tolerance is an engineering specification. It can be written down, tested against, and certified. It can also be written down badly, tested against cheaply, and certified by somebody with an interest in the answer.

How the Corps certifies, and what it costs

The Star Rangers Safety Corps maintains the strictest fold-transit standard in settled space. It is not, on the whole, admired for it. The Corps is conservative to the point of obstruction and slow in a way its own officers decline to apologise for, and the complaint filed against it by operators for six centuries has never really changed: that it treats every application as though the building were already on fire.

What it has, and what nothing else in settled space has, is a false-certification rate that rounds to zero. A hull the Safety Corps has passed is almost never a hull that should not have passed. Almost is load-bearing and the Corps is the party that insists on it — the certifications that failed are carried in its own loss-adjustment record, at length, with the reasoning that produced them intact. A body that files its own bad calls is a body whose good ones can be weighed.

The cost of that rigour is not hypothetical, and the Corps's critics are not wrong about it. A certification queue measured in years is a commercial fact, and an operator facing one has a lawful alternative in every direction: a registry whose standard is weaker and whose answer comes by return. The care that makes a Safety Corps certificate worth holding is the same care that sends tonnage to registries willing to issue one on request. The Corps has understood this for a century, has never found a way to move faster that survived its own review board, and regards the trade as settled rather than comfortable.

Two certificates, not one

A lawful transit requires two, issued by different bodies against different objects. The Safety Corps certifies the hull. The Navigation Corps certifies the route.

The division is not administrative tidiness. A fold route is not a fixed feature of anything — adjacency is a relation that holds for a while and then does not, and a corridor verified as safe is verified as of a date. The Navigation Corps' work is the continuous re-survey of drift routes: passages whose usable geometry moves, slowly and lawfully, away from the survey that certified them. A route certificate therefore carries an expiry in a way a hull certificate does not, and a great deal of the Corps' effort goes on the unglamorous business of establishing how fast a given corridor is drifting and how long that leaves.

This produces a third failure class, and the one that embarrasses everybody:

A properly certified hull on a lapsed route. Both certificates genuine, both issued by bodies that did their work correctly, and the loss occurs anyway because the corridor moved and nobody re-surveyed it in time. Nothing was evaded. The Navigation Corps files these as its own rather than the operator's wherever the re-survey interval was its to schedule, which is a practice its counterparts elsewhere have never adopted and its underwriters have never enjoyed.

The certification gap

The loss statistics are unambiguous about what the standards are worth: hulls and routes certified against them are lost at a rate the rest of the record does not approach.

The Safety Corps also has no power whatsoever to make anyone use it.

Certification is a matter for the polity registering the vessel. A great many polities have adopted the Safety Corps standard outright. Others maintain their own, written by bodies of varying independence, and a working minority of the fold-capable tonnage in settled space carries no meaningful certification at all — registered where registration is a fee rather than an inspection, and operating perfectly lawfully on that basis. The losses concentrate there with a consistency that has embarrassed everyone who has ever tried to argue it away.

Two further failure classes account for most of the remainder, and unlike the lapsed route, both involve something having been evaded:

Uncertified transit. No aperture tolerance was established, so none was exceeded. The Safety Corps' own phrasing for this in loss adjustment is that the vessel was operating outside a standard rather than against one, which is not a distinction its underwriters have ever accepted.
Certification against a weaker standard. More common, far harder to prosecute, and the reason the Safety Corps spends more of its budget on comparative standards work than on inspection. A hull can be fully, honestly, verifiably certified — to a standard whose aperture-stability margin the Safety Corps rejected two centuries ago as an economy taken at the wrong end.

The Hyperfold Yield Combine is the case the Corps teaches from, and it teaches from it because nothing in it was illegal. A licence, a counsel's opinion, an audited product. Lawful is not a synonym for safe, and the Safety Corps' whole institutional memory — a testing standard printed in Chicago in 1894 CE, written for insurers who wanted to know whether a thing would burn the building down — is organised around refusing to let the two words merge.

The Wrath of the Dragons

The devotional name attached first to a particular sequence of losses and has since been extended by ordinary usage to the whole class. It is the name most people in settled space actually reach for, and it long predates any of the standards work above. Since 2831 UCSD the event most people mean by it is Dock Seven, whose nine gallery witnesses supplied the word without being asked — and whatever answered the intrusion there, the record is plain about why there was an intrusion to answer: a licensed concern that the Rangers, holding no power to compel and built to hold none, could not make respect the safety of the fauna it was rendering.

The reading behind it is not naive, and the Archive's file is careful to say so. Dragons are low-ranking Levrils — a Cascade tier that genuinely acts, in Material time, on Material things. Agency is therefore available to the devotional account without any special pleading at all, which is exactly what makes the argument survive. A tradition holding that something with standing was angered, and that the losses followed, is making a claim its own cosmology licenses.

What the Institute published

The Institute produced the physical account, and — as with Dock Seven — it is the reading most of settled space actually holds, including a good many people who would describe themselves as devout.

The account requires no agency. Aperture instability, correlated with margin, correlated in turn with the standard a hull was built and certified against. In keeping with the house rule that makes Institute papers citable by people who despise the Institute, the publication carries what it failed to find: a residue of losses its model does not account for, listed at the same length as the finding, and left listed.

Its critics observe that the Institute has now published that residue nine times and concluded on each occasion that nothing requires explaining. The Institute's answer is that an unexplained remainder is not evidence of an author, and that the burden has not moved.

The third reading, and why the Archive keeps it

A body of scholarship disputes neither the physics nor the theology, but the transmission. What survives of the founding sequence is a literary tradition with recognisable conventions, compiled at a distance, by people with reasons — and on this reading the question is not what caused the losses but whether the sequence, as received, describes events that occurred in the order and the number given.

It is the least popular of the three positions and the one most often left out of the argument entirely. The Archive keeps it because it is the only one of the three that can be tested against something: the record itself, which is the Archive's own object.

All three are on file. None is adjudicated. What the file does record, and repeats in the margin of every revision, is the one finding all three readings are compatible with and none is required to explain: the losses fall where the certification is weakest. There are exceptions. They sit in the Safety Corps' own loss-adjustment record, they are few enough to be named individually, and every party to the argument has at some point tried to make one of them carry more weight than it will bear.