Maintain Two Approved Format-Bible Editions: Conditional Text or Separate Files?

Two approved editions of the same bible fail in opposite directions. A conditional source fails in the selection state: the native file still holds everything, the export does not, and the export looks finished. Separate files fail in the gap between them: each file reads correctly on its own, and the mistake exists only in the comparison nobody ran.

So the useful question is not which mechanism is safer in general. It is which of those two failures your team will actually catch. Pick the route whose failure your review step is built to find, then build the review step to check the whole edition rather than the passage you just touched. That second clause does most of the work in this article; the rest of it explains why.

One fact has to land before either route is worth discussing. Hidden conditional text stays in the document. In InDesign, a condition controls whether content is included in the flow, and hiding a passage does not remove it from the native file. The supplied reading of Adobe's Conditional text overview (updated 2 June 2026) records exactly that. A conditional source is therefore an edition-control tool, not a confidentiality tool. If you are hiding a passage because the recipient should not possess it, conditional text has not solved that problem — it has removed the passage from one view while leaving it in the file you are still handling. Keep that distinction separate from everything that follows, because the two ideas get merged constantly and they have nothing to do with each other.

Where the selection actually lives

The choice between one conditional source and two files is usually described as a trade between convenience and safety. That description is too coarse to act on. Both routes contain the same content; what differs is where the decision about which content appears is stored.

In a conditional source, that decision lives in a state you set: which named conditions are applied to which ranges, and which combination is currently shown. State is invisible in the document body. Two people can open the same file, see two different documents, and both believe they are looking at the file.

In separate files, the decision lives in the identity of the document. There is no state to misremember, because the selection was made when the file was created. That is a real advantage, and it costs you something specific: every shared passage now exists in more than one place, and no mechanism keeps those copies in agreement.

Neither arrangement is self-approving. A condition set decides what is included; it does not decide whether the wording is authorized. Two files decide nothing at all on that front. If a shared paragraph changes, both approved editions have been altered, and both approvals need re-confirmation. The conditional source saves you typing that change twice. It does not save you reviewing it twice.

A miniature, with its facts fixed

The bible below is invented to make the method checkable, and the files have not been built. What follows describes what to construct and what to look for, not results that exist.

Imagine a format bible with two approved editions: Edition P, written for the people deciding whether to commission the show, and Edition S, written for the crew who will build it. Most of the bible is identical in both. Three passages matter here.

S1 — shared, appears in both editions.

Round structure. Every episode runs four rounds. Rounds one to three are scored individually; round four is scored as a pair. A round ends when the host closes the board, and no answer given after the board closes counts.

A-P — the pitch alternative, approved for Edition P only.

The swap. This is the show's signature beat: two players trade the lead in the last ninety seconds, and the audience does not know who is ahead until the reveal.

A-S — the studio alternative, approved for Edition S only.

The swap. The host calls it at the board-closed cue. The score operator confirms both totals on the tally sheet before the host announces the exchange.

Q1 — a qualification required in both editions.

Approval note. The ninety-second swap window and the four-round count are fixed by the format owner. Any change to either requires written approval from the format owner before the change is used in a recorded episode.

The two alternatives are mutually appropriate, not ranked. A-P is not a softer version of A-S and A-S is not a hardened version of A-P; they answer different questions for different readers. Q1 is neither alternative nor shared description. It is a constraint that survives in both editions, and it is the passage most likely to be lost.

Failure one: a range one paragraph too long

In the conditional source, you name two conditions — call them PITCH and STUDIO — and save two visibility combinations as condition sets: Edition P shows PITCH and hides STUDIO; Edition S shows STUDIO and hides PITCH. S1 carries neither condition, so it stays in both. Q1 also carries neither condition, so it should stay in both.

Now apply the STUDIO condition to A-S, and drag the range one paragraph past the end of it. Q1 now carries STUDIO. Nothing has gone wrong that the file will tell you about. Set the source to Edition P and export: A-P appears, A-S is hidden, and Q1 has vanished, because Edition P hides STUDIO and Q1 is now STUDIO content.

The exported pitch edition is clean. It has no overset text, no missing font, no empty frame where the note used to be. It has simply lost a constraint that both editions are required to carry, and the readers of Edition P are the ones least likely to notice its absence, because they are not the people who will run the rounds. Reflow is a second consequence of the same cause: Q1 occupied space, and hiding it moves everything below it upward, so a literal page reference elsewhere in the pitch edition can land a paragraph or a page away from its target. Cross-references and section numbers survive reflow better than page numbers do, and either one is safer than a number typed into the sentence.

There is a mirror error, and it is worth knowing because teams often assume the opposite behavior. Conditions are not a restrictive ladder in which the most specific tag wins. Adobe's supplied documentation for Create and edit conditional text (2 June 2026) states that text remains visible when at least one applied condition is shown. So if A-P accidentally carries a leftover STUDIO tag from a duplicated draft, the studio edition will show both alternatives — and the reader will have two incompatible descriptions of the same beat with no indication of which one governs. The failure of an over-broad range removes necessary content; the failure of a stray second tag adds unauthorized content. Both are range mistakes, and both produce a document that looks deliberate.

Two habits follow. Apply conditions with the range boundaries visible, applying to the alternatives only and stopping before the qualification that follows. And look for override markers before exporting, because a saved condition set describes the combinations you defined, not a range someone nudged by hand afterward. A set that reads correctly can still produce a document carrying a manual exception.

Failure two: an update applied to one file

The separate-file route has no selection state, so failure one cannot happen. Instead: the format owner authorizes a change to S1, and the scoring split becomes rounds one and two individually scored, rounds three and four scored as a pair. The round count stays at four and the swap window is untouched, so Q1's terms are not triggered and Q1 does not need rewording. The change is applied to bible-S.indd. It is not applied to bible-P.indd.

Both files are internally consistent. The studio edition is correct and current. The pitch edition is correct as a document and wrong as an edition — it describes a scoring split the format no longer uses, in the same font, at the same level of confidence, with nothing anywhere in the file to suggest it is stale. No per-file inspection will find this. The only things that find it are a comparison between the two editions or a change log listing the shared edit with a tick against each file. If you keep neither, you have no defense, and the drift will be discovered by a reader.

The same update in the conditional source is one edit in S1 and both editions are current. That is the honest case for the conditional route, and it is narrower than it is usually made to sound: it saves the second edit, not the second review. Edition P was approved with a particular scoring split. Changing S1 changed an approved edition. Somebody still has to confirm that the pitch edition's approval extends to the new wording.

Choosing between them

Use the conditional source when the editions share most of their sentences, when the differences are a short list of passages someone can enumerate by name, when the native file lives in one place that nobody copies and edits offline, and when you will run a per-edition checklist against each export. Under those conditions the single source of truth is worth the state risk.

Use separate files when the editions diverge structurally — different sections, different order — because a condition set cannot express order that the source does not have, and when the divergence is broad enough that conditions would cover a large share of the text. Range boundaries multiply with coverage, and so does the chance of one being drawn a paragraph wrong. Separate files are also the better fit when more than one person edits, or when people work offline on copies, because there is no state for two people to disagree about. In exchange, you take on a mandatory shared-change log: every common edit gets a line, and every line gets a tick per file. Without that log, drift is not merely likely, it is undetectable until someone reads both editions side by side.

The build, in order

1. Write the edition definition before touching visibility. List S1 and everything like it, every alternative with its approved edition, and every qualification that must appear in both. Note the intended recipient and approval state of each. Approving a change is not part of this step. Neither is deciding which edition may carry which content; that decision is already made and this is implementation.
2. Apply named conditions to the alternatives only, leaving shared copy and qualifications outside the ranges. Save the visibility combinations as condition sets, one per edition.
3. Reconcile shared changes deliberately. In the conditional source, edit the shared passage once. In separate files, apply the same edit to each file and record the tick.
4. Set the source to one edition, check for override markers, and export. Then set the other edition and export again, checking markers each time.
5. Check each export against the edition definition — the full list, not just the passages you changed. Alternatives present and correct, qualification present, the other edition's alternative absent, no overset text, and page references landing where they claim to.
6. Reopen the native source with all conditions shown and confirm what it retains and where the ranges now fall. A clean visible edition says nothing about the source's contents.
7. Record the application version, the export settings, and which condition set produced each file. Fix the smallest bad range or unsynchronized passage rather than rebuilding, then recheck the outputs that range touched.

What to check in each output, and what to check in the source

"Verified" means two different checks here, and a record that conflates them is not usable. Visible content is what the reader receives: the right alternative, the right qualification, correct reflow, no overset. Retained content is what the native file holds with every condition shown: shared passages, both alternatives, and any older wording that was never removed.

The second check exists because the two are genuinely independent. A pitch export can be immaculate while the source contains the studio alternative, the studio wording of a passage, and a superseded draft. You already know that hidden text stays in the document; the operational consequence is that a source shared with anyone carries everything, regardless of how the export looks. If the source leaves your control, the edition control you built governs the export and nothing else.

One point of the source documentation is unsettled, and it should be recorded as unsettled rather than resolved. The two Adobe pages in hand do not agree about search: the overview page describes hidden conditional text as remaining editable and searchable in the document, while the create-and-edit page says hidden text is excluded from search. Which is accurate for your version and visibility state matters, because a searchable hidden passage can be found and navigated to by someone who does not know it is there. Do not carry an assumption either way into a workflow. Test search with conditions shown and with conditions hidden in the named application version, write down what you observe, and state the version alongside the observation. This uncertainty does not affect the argument above — hidden text is retained in the document regardless — but it does affect what you can honestly claim about who can find it.

Two editions, two records

The route you choose matters less than what you do after choosing it. Pick one, build both editions, and finish with two labeled files and two check records, each stating what is visible, what the source retains, which condition set or file produced it, and which application version and export settings were used. Keep visible and retained content as separate lines, because they are separate claims and only one of them is about what a reader sees.

Capture the evidence the record refers to, and capture the evidence that route actually produces. In a conditional source, that means the panel showing the range boundaries at the bad passage, the corrected inclusion state, and the exported result for each edition. With separate files, whose only state is which copy was edited, it means the diff between the two files or the change-log entry for the shared edit, alongside the exported result for each edition. Those captures are part of the deliverable, not decoration, and they are the difference between a workflow someone can audit and a workflow someone has to trust.

Frequently asked questions

Is hidden conditional text a way to keep content out of a recipient's hands?

No. Hidden conditional text stays in the document. In InDesign, a condition controls whether content is included in the flow, and hiding a passage does not remove it from the native file. Conditional text is an edition-control tool, not a confidentiality tool.

What is the likely failure with a conditional source?

The selection state can fail. If you apply a condition to an alternative and drag the range one paragraph too long, a qualification that should appear in both editions can become hidden and vanish from one export while the file still looks clean. A stray second tag can also make both alternatives visible, because text remains visible when at least one applied condition is shown. Apply with range boundaries visible, stop before shared qualifications, and check for override markers before exporting.

What is the likely failure with separate files?

A shared update applied to one file but not the other creates edition drift. Each file can be internally consistent and look correct, while one edition now describes an outdated rule. Per-file inspection will not find it; only a comparison between editions or a shared-change log with a tick per file will.

When should separate files be used instead of a conditional source?

Separate files fit when editions diverge structurally, including different sections or order, or when divergence is broad enough that conditions would cover a large share of text. They are also better when more than one person edits or people work offline on copies, because there is no state for two people to disagree about. In exchange, you need a mandatory shared-change log.

How should I verify the two editions?

Keep visible content and retained content as separate checks. Visible content is what the reader receives: the right alternative, the right qualification, correct reflow, no overset. Retained content is what the native file holds with all conditions shown. Record which condition set or file produced each output, the application version and export settings. Search behavior for hidden text is unsettled between the two Adobe pages cited, so test it in your version and record what you observe.

Further reading

• Build an InDesign Preflight Profile for a Treatment Handoff — For the part of your per-edition check you can automate, read how to build a scoped profile, test it with planted faults, and state its limits.
• Edit a Series Bible for the Collaborator Who Has to Build From It — Read before you write the edition definition: it shows how to mark which sentences govern a production decision, which merely illustrate, and which stay open.