Field Guide
Writing to ASD-STE100 in Practice
The Shape of It
A Dictionary and About Sixty-Five Rules
ASD-STE100 has two halves that work together. Part one is a set of writing rules, around sixty-five of them, covering words, noun phrases, verbs, sentence construction, procedures, descriptive writing, and safety instructions. Part two is a controlled dictionary in which each approved word carries one meaning and one part of speech.
That second half is the one people underestimate. English is generous with words that do several jobs, and technical documentation is where that generosity turns expensive. “ Service” can mean repair, lubricate, or supply. “ Follow” can mean obey or come after. “Check” can be a verb or a noun and means something different in each. A reader working in their second language, at night, on a stand, resolves that ambiguity with a guess. The dictionary removes the guess by assigning each word a single job and asking the writer to choose the precise one.
The audience explains the severity. Maintenance documentation is frequently read by people for whom English is not a first language, under time pressure, in conditions nobody would choose for careful reading. STE optimizes for exactly that reader, which is why its output can feel blunt to a writer and obvious to a technician. Those are the correct reactions in that order.
In Practice
What Changes in Your Sentences
Five rules do most of the work in most documents. Here they are with the kind of sentence they actually catch:
Remove the access panel and disconnect the electrical connector, then drain the reservoir before removing the pump.
1. Remove the access panel. 2. Disconnect the electrical connector. 3. Drain the reservoir. 4. Remove the pump.
A technician working with gloves and a torch reads one line at a time. Four actions in one sentence is four chances to lose your place.
The retaining bolts should be torqued to the specified value.
Torque the retaining bolts to 25 Nm.
Passive voice hides who acts. It also hid the value, which is the only part the reader needed.
Follow the procedure to service the unit before it is released.
Obey the procedure to lubricate the unit before you release it.
“Follow” can mean obey or come after. “Service” can mean repair, lubricate, or supply. STE assigns one meaning per word and the writer picks the precise one.
Removing the cover, examine the seal for damage occurring during operation.
Remove the cover. Examine the seal for damage.
Participles are a common source of ambiguity for non-native readers, and most of them disappear without loss.
If the indication does not illuminate after the test switch has been pressed and held for the required interval, it is possible that the lamp has failed.
Press and hold the test switch for 3 seconds. If the indication does not come on, the lamp is defective.
The rules cap procedural sentences at 20 words and descriptive sentences at 25. The cap is a symptom check, not the goal.
Where Teams Get It Wrong
Three Predictable Mistakes
Treating the checker as the standard. Software enforces the mechanical rules well and cannot evaluate whether the instruction is right, whether the sequence works, or whether the approved word you chose is the correct one here. Passing the checker is a floor, in precisely the way that passing schema validation is a floor rather than proof that the content is any good.
Retrofitting without pricing re-verification. Rewriting approved content changes approved text, and in regulated documentation changed text can trigger re-verification. That converts an editorial project into a certification one, which is a different budget and a different approver. The workable pattern is to apply STE from a fixed date to new and substantially revised content, tolerate a mixed library for a while, and convert the rest when it is being touched anyway.
Leaving the terminology uncontrolled. STE governs general vocabulary, but your product's own technical names are yours to manage, and a controlled general vocabulary sitting on top of eleven names for the same assembly achieves very little. That is a defined terms problem, solved with a managed list rather than a style note, and it belongs in the same program as the STE adoption rather than after it. The same logic applies to safety messages, where a sentence the reader has to parse twice is least affordable.
The Business Case
It Pays for Itself in Translation
Comprehension is the stated reason for STE and it is genuinely hard to measure, which makes it a weak argument in a budget meeting. There is a stronger one available and it is rarely made. Controlled source language reduces translation cost. Shorter sentences, consistent terminology, and one meaning per word produce higher translation memory match rates, less variance between translators, and fewer queries coming back. For an organization publishing into several languages, the localization saving is measurable in a way that comprehension is not, and it usually arrives first.
There is a second beneficiary nobody was designing for. Consistent, unambiguous source text is also what makes content tractable for retrieval and machine reading. A library where one concept has one name and one sentence carries one instruction is a library a model can ground on. STE was written decades before anyone needed that, for a human reader in bad light, and the discipline turns out to serve both.
FAQ
Questions We Hear
What is ASD-STE100?
A controlled language specification for technical documentation, made of two parts: a set of writing rules, about sixty-five of them, and a dictionary of approved words where each entry carries one meaning and one part of speech. It began in civil aviation maintenance documentation and is now specified or recommended well beyond it, across aerospace and defense and increasingly in other regulated industries. The current issue is Issue 9, published in January 2025, and the specification is issued free of charge on request through the ASD Simplified Technical English Maintenance Group.
Does Simplified Technical English mean dumbing down the content?
No, and the objection usually comes from writers rather than readers. STE constrains vocabulary and sentence construction, not technical precision. Technical names and technical verbs specific to your product remain available, and the dictionary exists to stop one word carrying three meanings rather than to remove words. The result reads plainer, which is the intent: the audience frequently includes people reading in their second or third language, under time pressure, in poor light, wearing gloves.
Do we need STE checker software?
For anything at scale, yes, though it is worth being clear about what it does. A checker enforces the mechanical rules reliably: unapproved words, sentence length, banned constructions, and dictionary compliance. What it cannot do is tell you whether the instruction is correct, whether the steps are in a workable order, or whether the approved word you chose is the right one for this situation. Treat it the way you treat a schema validator, as a floor rather than a standard, and keep an editorial pass above it.
Can we retrofit STE to existing manuals?
You can, and the cost is usually underestimated for a reason that has nothing to do with writing. Rewriting content that has been approved changes approved text, and in regulated documentation changed text can require re-verification. That turns an editorial project into a certification project, and the budget stops being about words. The workable pattern is to apply STE to new and substantially revised content from a fixed date, accept a mixed library for a period, and convert the rest only when it is being revised anyway.
Is STE mandatory?
Only when something makes it so, and that something is normally a contract or a program's writing standard rather than a regulator. Its reach is wider than its citations, though, because many organizations adopt it voluntarily once they see the translation savings. That is the underrated commercial argument: controlled source language reduces translation cost and variance across every target language, so for anyone publishing in more than a few languages STE tends to pay for itself in localization before anyone measures comprehension.
Get In Touch
Adopting STE?
We train writers into it, set the terminology management that has to run alongside it, and are honest about what retrofitting existing approved content will cost you.