Technology readiness, open questions and responsible engineering research

State the claim and the comparison

Describe the intended function, environment, scale, duration and relevant users or stakeholders. Distinguish an observed result, a model prediction and a future target. A comparison should use compatible service and boundaries; an impressive component number can otherwise hide the burdens moved elsewhere in the system.

Retain the method, configuration, uncertainty and unsuccessful results. Independent reproduction and investigation of discrepancies provide information that a selected demonstration cannot. A result should be traceable to the exact specimen, software and conditions that produced it.

Read a readiness framework in its context

NASA technology-readiness frameworks distinguish evidence from principles and laboratory demonstrations through increasingly representative system and operational demonstrations. The relevant environment is defined around the intended application and its critical conditions. Framework details vary by program, so identify the definitions being used.

A mature component does not automatically make an integrated system mature. Interfaces, scale, support equipment, failure behaviour and the operating context may introduce new questions. Averaging several component readiness numbers creates no justified system rating, and taking the lowest number is not a substitute for integration evidence either.

Maturity is also different from superiority or acceptability. A demonstrated technology can still be costly, difficult to maintain or unsuitable for a particular need. Conversely, an early idea can be scientifically valuable while its service feasibility remains open.

Choose evidence that can change a specific claim

Give the earlier twenty-condition plan a concrete, invented setting: the research finger fixture just described, configuration C1, including its element, linkage, sensor calibration and controller R1. Eighteen distinct bench conditions combine three commanded displacements (2, 4 and 6 mm), three opposing loads (1, 2 and 3 N) and two recorded starting histories (cooled and recently actuated), all at 20 °C on laboratory host H1. These are supplied research conditions, not operating recommendations. For each condition, require the absolute position error to remain at most 0.5 mm throughout the interval from 3.0 to 4.0 s after the command. Supply checked logs, with measurement limits accounted for, showing that all eighteen met that criterion.

The two remaining conditions answer different questions. W uses the 6 mm command, 3 N load and recently actuated history at 35 °C on H1: does the change in thermal surroundings invalidate that particular motion claim? I uses the 6 mm command, 3 N load and cooled history at 20 °C on intended host H2, including its command and sensor-data connection: does the integrated interface preserve the function? Temperature matters because heat exchange and transformation history influence motion. The host matters because changed units, timing or message handling can change what the controller sees even when the mechanical element is unchanged.

The present positive conclusion is that C1 met the supplied criterion in those eighteen H1 bench conditions. That is evidence for a bounded laboratory function, not merely a successful-looking motion. To investigate the claimed thermal extension, W is the direct next comparison: retain C1 and the identified load, command and history while changing the stated ambient condition. I separately needs an end-to-end record tying transmitted commands, position units and timestamps to the actual motion on H2. Neither result substitutes for the other. This example assigns no readiness number and establishes no continuous temperature/load range from a few named conditions.

Make unknowns useful

List the unresolved questions that could change the decision and identify the evidence needed to answer them. Prioritize by consequence and uncertainty, not only by which test is easiest to perform. A plan is useful evidence of intended work; it is not evidence that the result has already occurred.

Responsible research considers who bears risks and costs, who benefits, resource and environmental burdens, and how changes will be governed. Preserve limitations when communicating progress. A clear remaining question is more useful to the next engineer than an unsupported claim that development is complete.

Separate observed passes from the untested plan

Construct a plan containing 20 distinct conditions. Eighteen have been executed and all eighteen passed their supplied criteria. Two have not been executed. The observed pass fraction among executed conditions is 18/18=100%, while the simple count-based plan coverage is 18/20=90%.

The two missing outcomes are unknown, not inferred passes or observed failures. If they represent particularly important conditions, a count-based 90% coverage figure may also conceal their importance. The figure preserves their separate status.

These 20 conditions are not stated to be independent random trials of an operating population. The fractions establish neither a reliability probability nor a readiness level. A complete claim would need the importance and representativeness of the conditions, valid criteria, uncertainty, configuration evidence and the relevant integrated demonstration.

For the filled study, 3 commands × 3 loads × 2 starting histories gives the eighteen executed conditions. The two named additions make the total twenty. The 90% count coverage remains correct, but now its missing 10% has a meaning: one thermal extension and one host/interface comparison. Reporting the supported C1/H1 bench result while keeping those claims open is a useful research conclusion. It directs the next evidence without turning plan coverage into reliability or a universal readiness formula.

Try the reasoning

1. Can 100% of executed tests passing establish full planned coverage?

   Show the answer and reasoning

   No. Two of the twenty planned conditions remain unexecuted in the example.

2. Is system readiness the arithmetic mean of component ratings?

   Show the answer and reasoning

   No. The integrated system and intended conditions require their own evidence argument.

3. What distinguishes a target from a result?

   Show the answer and reasoning

   A target states intended performance; a result is an observation or prediction with a stated method and evidence scope.

4. A new checked W record, for the exact specified C1/H1 condition, shows an absolute position error no greater than 0.4 mm throughout 3.0–4.0 s, with measurement limits already accounted for. I remains unexecuted. What can now be concluded, and what should the next evidence examine?

   Show the answer and reasoning

   W meets the supplied 0.5 mm criterion for that condition, so the record adds positive evidence for the named 35 °C case. There are now 19 observed passes out of 19 executed conditions and 19/20 = 95% count coverage. It does not establish every temperature, history or load between the tested cases. I still needs the H2 command/sensor connection and actual motion assessed together, preserving position units, timestamps and configuration. Closing W does not establish that the intended host interprets or delivers the same feedback correctly.

What to carry forward

Keep mechanisms, demonstrations, open questions and lifecycle consequences visible as technology advances.