Topic · non functional requirements

Non functional requirements

Gist

Non functional requirements on this hop are five authored categories on the requirement set, not an ISO 25010 score. Proof runs proof status. Wikipedia still owns the glossary. Jama still authors.

proof status

Keep SAFe if you already write NFRs as solution constraints. Keep ISO 25010 if you already score quality attributes. Proof will not classify a shall as FR or NFR for you.

01 · The missed spine

Every shall is functional. The runtime has no bound.

The category table is a smell test. It is not a quota.

Wikipedia, GeeksforGeeks, Jama, and SAFe will give you non functional requirements: performance, security, reliability, maintainability. That is not this URL. Proof's slice prints how many live shalls sit in functional, performance, constraint, interface, and safety. The order is stable so a diff is a real change. A small library may be mostly functional. A parser, a runtime, or a security boundary that is 100% functional is the miss.

The motivating miss is mechanical. The feature shalls shipped. Latency, resource use, forbidden states, and the wire contract never got a row. Coverage and the test runner are green. proof status still shows one category. The next agent treats the glossary page as proof the quality attributes exist. They do not, as files.

Three honest exits, in order: author the missing category on a real shall; set --category on a shall that already said the bound; or keep going when the component really is one functional job. Relabelling every shall to constraint to make the table look mixed is how the number went green while the bound stayed unwritten.

proof status
proof req set SYS-REQ-010 --category performance
proof help category-balance

The first command is the table. The set is a real category, not a comment. The help topic is the smell test. A green ISO 25010 worksheet next to 40 functional shalls is how the runtime stayed unbounded.

02 · The exhibit

Same repo. A quality list, or this table.

The glossary is complete. The spec set is not. Click the tabs.

The run

  • Ask Wikipedia / SAFe / ISO 25010 / Jama NFR guide
  • Exit 0 if the list names performance and security
  • Why the list does not count files
List green

This hop

No category table. Every live shall is still functional.

No table

The run

Keep the list. It still owns the glossary. That is not this hop.

Keep the glossary

Proof

  • Ask functional, performance, constraint, interface, safety
  • Out a stable count per category. A nudge, never a gate
A green list is not a table

Same repo. A quality list, or this table. Click the tabs.

Surface What they do What Proof does What we lose
Wikipedia / GeeksforGeeks Define NFR as criteria for how the system operates. Count authored categories on live shalls. Not a glossary. Keep Wikipedia if you need the term.
ISO 25010 A quality model: eight characteristics, subcharacteristics, measures. Five categories. No score. Not ISO 25010. We have not scored this slice against that model.
SAFe NFRs as solution constraints that guide design. A status table. Author the bound as a shall if it is missing. Not SAFe. Keep the constraint list if you already write it there.
Jama The authoring programme. NFR types if you put them there. Print the mix. Set --category on one file. Not Jama's type field. Jama still authors.
category table A quota, or an FR/NFR classifier. A smell test. A small library may be mostly functional. The table never blocks a stage. Relabelling is not a bound.

The teaching graph is still a complete quality list next to a spec that is 100% functional. Author the latency shall if the runtime has a bound. Set the category if the bound is already in the file. Keep the mix when the job is one functional pile. Do not retag folders to make the table look mixed.

proof status
proof status --format json
proof req set SYS-REQ-010 --category constraint

The check is a status line. It never fails a stage. The known categories are fixed: functional, performance, constraint, interface, safety. Other names sort after those. See requirements quality if the missing object is rationale on a shall. See software metrics if the missing object is component count, not category mix. See hazard analysis if the missing object is a hazard control, not a safety label. See Proof vs Jama if the missing object is the authoring programme.

03 · The honest loss

Proof names the mix. It does not write the bound.

A category table is not a proof of the Go. Jama still authors.

Proof does not implement ISO 25010, SAFe NFR worksheets, or an FR/NFR classifier. It does not score quality attributes. A 100% functional mix is a nudge, not a failed gate. A library that legitimately has no performance shall should keep reporting one category. We have not scored this slice against a frozen ISO 25010 or SAFe corpus. The loss is named, not scored. Retagging every shall to look mixed converts a visible gap into an invisible claim.

Rationale on a shall stays on requirements quality. Component count stays on software metrics. Jama still authors.

04 · Nearby questions

What people type next.

What are non functional requirements? Same question. Same URL.

What are quality attributes? Same cluster in Proof. The Google result is often ISO 25010. That is not this hop.

What is ISO 25010? A quality model. Proof has no score for it.

What is software quality? Delivery and defect density live on software metrics. Not this hop.

Is Proof a Jama alternative for NFRs? No. Jama still authors. Proof vs Jama.

Is Proof a SAFe NFR tool? No. Keep the constraint list if you already write it there.